Merge branch 'basecamp:main' into buildpacks

lib/kamal/cli.rb

@@ -1,4 +1,5 @@
 module Kamal::Cli
+  class BootError < StandardError; end
   class HookError < StandardError; end
   class LockError < StandardError; end
 end

lib/kamal/cli/accessory.rb

@@ -1,16 +1,17 @@
 class Kamal::Cli::Accessory < Kamal::Cli::Base
   desc "boot [NAME]", "Boot new accessory service on host (use NAME=all to boot all accessories)"
-  def boot(name, login: true)
+  def boot(name, prepare: true)
     with_lock do
       if name == "all"
         KAMAL.accessory_names.each { |accessory_name| boot(accessory_name) }
       else
+        prepare(name) if prepare
+
         with_accessory(name) do |accessory, hosts|
           directories(name)
           upload(name)
 
           on(hosts) do
-            execute *KAMAL.registry.login if login
             execute *KAMAL.auditor.record("Booted #{name} accessory"), verbosity: :debug
             execute *accessory.ensure_env_directory
             upload! accessory.secrets_io, accessory.secrets_path, mode: "0600"
@@ -57,15 +58,10 @@ class Kamal::Cli::Accessory < Kamal::Cli::Base
       if name == "all"
         KAMAL.accessory_names.each { |accessory_name| reboot(accessory_name) }
       else
-        with_accessory(name) do |accessory, hosts|
-          on(hosts) do
-            execute *KAMAL.registry.login
-          end
-
-          stop(name)
-          remove_container(name)
-          boot(name, login: false)
-        end
+        prepare(name)
+        stop(name)
+        remove_container(name)
+        boot(name, prepare: false)
       end
     end
   end
@@ -97,10 +93,8 @@ class Kamal::Cli::Accessory < Kamal::Cli::Base
   desc "restart [NAME]", "Restart existing accessory container on host"
   def restart(name)
     with_lock do
-      with_accessory(name) do
-        stop(name)
-        start(name)
-      end
+      stop(name)
+      start(name)
     end
   end
 
@@ -153,23 +147,25 @@ class Kamal::Cli::Accessory < Kamal::Cli::Base
   option :grep, aliases: "-g", desc: "Show lines with grep match only (use this to fetch specific requests by id)"
   option :grep_options, aliases: "-o", desc: "Additional options supplied to grep"
   option :follow, aliases: "-f", desc: "Follow logs on primary server (or specific host set by --hosts)"
+  option :skip_timestamps, type: :boolean, aliases: "-T", desc: "Skip appending timestamps to logging output"
   def logs(name)
     with_accessory(name) do |accessory, hosts|
       grep = options[:grep]
       grep_options = options[:grep_options]
+      timestamps = !options[:skip_timestamps]
 
       if options[:follow]
         run_locally do
           info "Following logs on #{hosts}..."
-          info accessory.follow_logs(grep: grep, grep_options: grep_options)
-          exec accessory.follow_logs(grep: grep, grep_options: grep_options)
+          info accessory.follow_logs(timestamps: timestamps, grep: grep, grep_options: grep_options)
+          exec accessory.follow_logs(timestamps: timestamps, grep: grep, grep_options: grep_options)
         end
       else
         since = options[:since]
         lines = options[:lines].presence || ((since || grep) ? nil : 100) # Default to 100 lines if since or grep isn't set
 
         on(hosts) do
-          puts capture_with_info(*accessory.logs(since: since, lines: lines, grep: grep, grep_options: grep_options))
+          puts capture_with_info(*accessory.logs(timestamps: timestamps, since: since, lines: lines, grep: grep, grep_options: grep_options))
         end
       end
     end
@@ -224,6 +220,25 @@ class Kamal::Cli::Accessory < Kamal::Cli::Base
     end
   end
 
+  desc "upgrade", "Upgrade accessories from Kamal 1.x to 2.0 (restart them in 'kamal' network)"
+  option :rolling, type: :boolean, default: false, desc: "Upgrade one host at a time"
+  option :confirmed, aliases: "-y", type: :boolean, default: false, desc: "Proceed without confirmation question"
+  def upgrade(name)
+    confirming "This will restart all accessories" do
+      with_lock do
+        host_groups = options[:rolling] ? KAMAL.accessory_hosts : [ KAMAL.accessory_hosts ]
+        host_groups.each do |hosts|
+          host_list = Array(hosts).join(",")
+          KAMAL.with_specific_hosts(hosts) do
+            say "Upgrading #{name} accessories on #{host_list}...", :magenta
+            reboot name
+            say "Upgraded #{name} accessories on #{host_list}...", :magenta
+          end
+        end
+      end
+    end
+  end
+
   private
     def with_accessory(name)
       if KAMAL.config.accessory(name)
@@ -251,11 +266,20 @@ class Kamal::Cli::Accessory < Kamal::Cli::Base
     end
 
     def remove_accessory(name)
-      with_accessory(name) do
-        stop(name)
-        remove_container(name)
-        remove_image(name)
-        remove_service_directory(name)
+      stop(name)
+      remove_container(name)
+      remove_image(name)
+      remove_service_directory(name)
+    end
+
+    def prepare(name)
+      with_accessory(name) do |accessory, hosts|
+        on(hosts) do
+          execute *KAMAL.registry.login
+          execute *KAMAL.docker.create_network
+        rescue SSHKit::Command::Failed => e
+          raise unless e.message.include?("already exists")
+        end
       end
     end
 end

lib/kamal/cli/app.rb

@@ -4,7 +4,7 @@ class Kamal::Cli::App < Kamal::Cli::Base
     with_lock do
       say "Get most recent version available as an image...", :magenta unless options[:version]
       using_version(version_or_latest) do |version|
-        say "Start container with version #{version} using a #{KAMAL.config.readiness_delay}s readiness delay (or reboot if already running)...", :magenta
+        say "Start container with version #{version} (or reboot if already running)...", :magenta
 
         # Assets are prepared in a separate step to ensure they are on all hosts before booting
         on(KAMAL.hosts) do
@@ -38,8 +38,17 @@ class Kamal::Cli::App < Kamal::Cli::Base
         roles = KAMAL.roles_on(host)
 
         roles.each do |role|
+          app = KAMAL.app(role: role, host: host)
           execute *KAMAL.auditor.record("Started app version #{KAMAL.config.version}"), verbosity: :debug
-          execute *KAMAL.app(role: role, host: host).start, raise_on_non_zero_exit: false
+          execute *app.start, raise_on_non_zero_exit: false
+
+          if role.running_proxy?
+            version = capture_with_info(*app.current_running_version, raise_on_non_zero_exit: false).strip
+            endpoint = capture_with_info(*app.container_id_for_version(version)).strip
+            raise Kamal::Cli::BootError, "Failed to get endpoint for #{role} on #{host}, did the container boot?" if endpoint.empty?
+
+            execute *app.deploy(target: endpoint)
+          end
         end
       end
     end
@@ -52,8 +61,18 @@ class Kamal::Cli::App < Kamal::Cli::Base
         roles = KAMAL.roles_on(host)
 
         roles.each do |role|
+          app = KAMAL.app(role: role, host: host)
           execute *KAMAL.auditor.record("Stopped app", role: role), verbosity: :debug
-          execute *KAMAL.app(role: role, host: host).stop, raise_on_non_zero_exit: false
+
+          if role.running_proxy?
+            version = capture_with_info(*app.current_running_version, raise_on_non_zero_exit: false).strip
+            endpoint = capture_with_info(*app.container_id_for_version(version)).strip
+            if endpoint.present?
+              execute *app.remove(target: endpoint), raise_on_non_zero_exit: false
+            end
+          end
+
+          execute *app.stop, raise_on_non_zero_exit: false
         end
       end
     end
@@ -169,12 +188,14 @@ class Kamal::Cli::App < Kamal::Cli::Base
   option :grep, aliases: "-g", desc: "Show lines with grep match only (use this to fetch specific requests by id)"
   option :grep_options, aliases: "-o", desc: "Additional options supplied to grep"
   option :follow, aliases: "-f", desc: "Follow log on primary server (or specific host set by --hosts)"
+  option :skip_timestamps, type: :boolean, aliases: "-T", desc: "Skip appending timestamps to logging output"
   def logs
     # FIXME: Catch when app containers aren't running
 
     grep = options[:grep]
     grep_options = options[:grep_options]
     since = options[:since]
+    timestamps = !options[:skip_timestamps]
 
     if options[:follow]
       lines = options[:lines].presence || ((since || grep) ? nil : 10) # Default to 10 lines if since or grep isn't set
@@ -186,8 +207,8 @@ class Kamal::Cli::App < Kamal::Cli::Base
         role = KAMAL.roles_on(KAMAL.primary_host).first
 
         app = KAMAL.app(role: role, host: host)
-        info app.follow_logs(host: KAMAL.primary_host, lines: lines, grep: grep, grep_options: grep_options)
-        exec app.follow_logs(host: KAMAL.primary_host, lines: lines, grep: grep, grep_options: grep_options)
+        info app.follow_logs(host: KAMAL.primary_host, timestamps: timestamps, lines: lines, grep: grep, grep_options: grep_options)
+        exec app.follow_logs(host: KAMAL.primary_host, timestamps: timestamps, lines: lines, grep: grep, grep_options: grep_options)
       end
     else
       lines = options[:lines].presence || ((since || grep) ? nil : 100) # Default to 100 lines if since or grep isn't set
@@ -197,7 +218,7 @@ class Kamal::Cli::App < Kamal::Cli::Base
 
         roles.each do |role|
           begin
-            puts_by_host host, capture_with_info(*KAMAL.app(role: role, host: host).logs(since: since, lines: lines, grep: grep, grep_options: grep_options))
+            puts_by_host host, capture_with_info(*KAMAL.app(role: role, host: host).logs(timestamps: timestamps, since: since, lines: lines, grep: grep, grep_options: grep_options))
           rescue SSHKit::Command::Failed
             puts_by_host host, "Nothing found"
           end
@@ -212,6 +233,7 @@ class Kamal::Cli::App < Kamal::Cli::Base
       stop
       remove_containers
       remove_images
+      remove_app_directory
     end
   end
 
@@ -253,6 +275,20 @@ class Kamal::Cli::App < Kamal::Cli::Base
     end
   end
 
+  desc "remove_app_directory", "Remove the service directory from servers", hide: true
+  def remove_app_directory
+    with_lock do
+      on(KAMAL.hosts) do |host|
+        roles = KAMAL.roles_on(host)
+
+        roles.each do |role|
+          execute *KAMAL.auditor.record("Removed #{KAMAL.config.app_directory} on all servers", role: role), verbosity: :debug
+          execute *KAMAL.server.remove_app_directory, raise_on_non_zero_exit: false
+        end
+      end
+    end
+  end
+
   desc "version", "Show app version currently running on servers"
   def version
     on(KAMAL.hosts) do |host|

lib/kamal/cli/app/boot.rb

@@ -1,7 +1,7 @@
 class Kamal::Cli::App::Boot
   attr_reader :host, :role, :version, :barrier, :sshkit
   delegate :execute, :capture_with_info, :capture_with_pretty_json, :info, :error, :upload!, to: :sshkit
-  delegate :uses_cord?, :assets?, :running_traefik?, to: :role
+  delegate :assets?, :running_proxy?, to: :role
 
   def initialize(host, role, sshkit, version, barrier)
     @host = host
@@ -45,15 +45,22 @@ class Kamal::Cli::App::Boot
 
     def start_new_version
       audit "Booted app version #{version}"
-
-      execute *app.tie_cord(role.cord_host_file) if uses_cord?
       hostname = "#{host.to_s[0...51].gsub(/\.+$/, '')}-#{SecureRandom.hex(6)}"
 
       execute *app.ensure_env_directory
       upload! role.secrets_io(host), role.secrets_path, mode: "0600"
-      execute *app.run(hostname: hostname)
 
-      Kamal::Cli::Healthcheck::Poller.wait_for_healthy(pause_after_ready: true) { capture_with_info(*app.status(version: version)) }
+      execute *app.run(hostname: hostname)
+      if running_proxy?
+        endpoint = capture_with_info(*app.container_id_for_version(version)).strip
+        raise Kamal::Cli::BootError, "Failed to get endpoint for #{role} on #{host}, did the container boot?" if endpoint.empty?
+        execute *app.deploy(target: endpoint)
+      else
+        Kamal::Cli::Healthcheck::Poller.wait_for_healthy(pause_after_ready: true) { capture_with_info(*app.status(version: version)) }
+      end
+    rescue => e
+      error "Failed to boot #{role} on #{host}"
+      raise e
     end
 
     def stop_new_version
@@ -61,16 +68,7 @@ class Kamal::Cli::App::Boot
     end
 
     def stop_old_version(version)
-      if uses_cord?
-        cord = capture_with_info(*app.cord(version: version), raise_on_non_zero_exit: false).strip
-        if cord.present?
-          execute *app.cut_cord(cord)
-          Kamal::Cli::Healthcheck::Poller.wait_for_unhealthy(pause_after_ready: true) { capture_with_info(*app.status(version: version)) }
-        end
-      end
-
       execute *app.stop(version: version), raise_on_non_zero_exit: false
-
       execute *app.clean_up_assets if assets?
     end
 

lib/kamal/cli/base.rb

@@ -101,7 +101,7 @@ module Kamal::Cli
       end
 
       def acquire_lock
-        ensure_run_and_locks_directory
+        ensure_run_directory
 
         raise_if_locked do
           say "Acquiring the deploy lock...", :magenta
@@ -135,8 +135,10 @@ module Kamal::Cli
           details = { hosts: KAMAL.hosts.join(","), command: command, subcommand: subcommand }
 
           say "Running the #{hook} hook...", :magenta
-          run_locally do
-            execute *KAMAL.hook.run(hook, **details, **extra_details)
+          with_env KAMAL.hook.env(**details, **extra_details) do
+            run_locally do
+              execute *KAMAL.hook.run(hook)
+            end
           rescue SSHKit::Command::Failed => e
             raise HookError.new("Hook `#{hook}` failed:\n#{e.message}")
           end
@@ -174,14 +176,23 @@ module Kamal::Cli
         instance_variable_get("@_invocations").first
       end
 
-      def ensure_run_and_locks_directory
+      def reset_invocation(cli_class)
+        instance_variable_get("@_invocations")[cli_class].pop
+      end
+
+      def ensure_run_directory
         on(KAMAL.hosts) do
           execute(*KAMAL.server.ensure_run_directory)
         end
+      end
 
-        on(KAMAL.primary_host) do
-          execute(*KAMAL.lock.ensure_locks_directory)
-        end
+      def with_env(env)
+        current_env = ENV.to_h.dup
+        ENV.update(env)
+        yield
+      ensure
+        ENV.clear
+        ENV.update(current_env)
       end
   end
 end

lib/kamal/cli/build.rb

@@ -30,28 +30,30 @@ class Kamal::Cli::Build < Kamal::Cli::Base
       say "Building with uncommitted changes:\n #{uncommitted_changes}", :yellow
     end
 
-    run_locally do
-      begin
-        execute *KAMAL.builder.inspect_builder
-      rescue SSHKit::Command::Failed => e
-        if e.message =~ /(context not found|no builder|no compatible builder|does not exist)/
-          warn "Missing compatible builder, so creating a new one first"
-          begin
-            cli.remove
-          rescue SSHKit::Command::Failed
-            raise unless e.message =~ /(context not found|no builder|does not exist)/
+    with_env(KAMAL.config.builder.secrets) do
+      run_locally do
+        begin
+          execute *KAMAL.builder.inspect_builder
+        rescue SSHKit::Command::Failed => e
+          if e.message =~ /(context not found|no builder|no compatible builder|does not exist)/
+            warn "Missing compatible builder, so creating a new one first"
+            begin
+              cli.remove
+            rescue SSHKit::Command::Failed
+              raise unless e.message =~ /(context not found|no builder|does not exist)/
+            end
+            cli.create
+          else
+            raise
           end
-          cli.create
-        else
-          raise
         end
-      end
 
-      # Get the command here to ensure the Dir.chdir doesn't interfere with it
-      push = KAMAL.builder.push
+        # Get the command here to ensure the Dir.chdir doesn't interfere with it
+        push = KAMAL.builder.push
 
-      KAMAL.with_verbosity(:debug) do
-        Dir.chdir(KAMAL.config.builder.build_directory) { execute *push, env: KAMAL.config.builder.secrets }
+        KAMAL.with_verbosity(:debug) do
+          Dir.chdir(KAMAL.config.builder.build_directory) { execute *push }
+        end
       end
     end
   end

lib/kamal/cli/healthcheck/poller.rb

@@ -1,26 +1,30 @@
 module Kamal::Cli::Healthcheck::Poller
   extend self
 
-  TRAEFIK_UPDATE_DELAY = 5
-
-
-  def wait_for_healthy(pause_after_ready: false, &block)
+  def wait_for_healthy(role, &block)
     attempt = 1
-    max_attempts = KAMAL.config.healthcheck.max_attempts
+    timeout_at = Time.now + KAMAL.config.deploy_timeout
+    readiness_delay = KAMAL.config.readiness_delay
 
     begin
-      case status = block.call
-      when "healthy"
-        sleep TRAEFIK_UPDATE_DELAY if pause_after_ready
-      when "running" # No health check configured
-        sleep KAMAL.config.readiness_delay if pause_after_ready
-      else
-        raise Kamal::Cli::Healthcheck::Error, "container not ready (#{status})"
+      status = block.call
+
+      if status == "running"
+        # Wait for the readiness delay and confirm it is still running
+        if readiness_delay > 0
+          info "Container is running, waiting for readiness delay of #{readiness_delay} seconds"
+          sleep readiness_delay
+          status = block.call
+        end
+      end
+
+      unless %w[ running healthy ].include?(status)
+        raise Kamal::Cli::Healthcheck::Error, "container not ready after #{KAMAL.config.deploy_timeout} seconds (#{status})"
       end
     rescue Kamal::Cli::Healthcheck::Error => e
-      if attempt <= max_attempts
-        info "#{e.message}, retrying in #{attempt}s (attempt #{attempt}/#{max_attempts})..."
-        sleep attempt
+      time_left = timeout_at - Time.now
+      if time_left > 0
+        sleep [ attempt, time_left ].min
         attempt += 1
         retry
       else
@@ -31,31 +35,6 @@ module Kamal::Cli::Healthcheck::Poller
     info "Container is healthy!"
   end
 
-  def wait_for_unhealthy(pause_after_ready: false, &block)
-    attempt = 1
-    max_attempts = KAMAL.config.healthcheck.max_attempts
-
-    begin
-      case status = block.call
-      when "unhealthy"
-        sleep TRAEFIK_UPDATE_DELAY if pause_after_ready
-      else
-        raise Kamal::Cli::Healthcheck::Error, "container not unhealthy (#{status})"
-      end
-    rescue Kamal::Cli::Healthcheck::Error => e
-      if attempt <= max_attempts
-        info "#{e.message}, retrying in #{attempt}s (attempt #{attempt}/#{max_attempts})..."
-        sleep attempt
-        attempt += 1
-        retry
-      else
-        raise
-      end
-    end
-
-    info "Container is unhealthy!"
-  end
-
   private
     def info(message)
       SSHKit.config.output.info(message)

lib/kamal/cli/lock.rb

@@ -12,7 +12,7 @@ class Kamal::Cli::Lock < Kamal::Cli::Base
   option :message, aliases: "-m", type: :string, desc: "A lock message", required: true
   def acquire
     message = options[:message]
-    ensure_run_and_locks_directory
+    ensure_run_directory
 
     raise_if_locked do
       on(KAMAL.primary_host) do

lib/kamal/cli/main.rb

@@ -35,8 +35,8 @@ class Kamal::Cli::Main < Kamal::Cli::Base
       with_lock do
         run_hook "pre-deploy", secrets: true
 
-        say "Ensure Traefik is running...", :magenta
-        invoke "kamal:cli:traefik:boot", [], invoke_options
+        say "Ensure kamal-proxy is running...", :magenta
+        invoke "kamal:cli:proxy:boot", [], invoke_options
 
         say "Detect stale containers...", :magenta
         invoke "kamal:cli:app:stale_containers", [], invoke_options.merge(stop: true)
@@ -48,10 +48,10 @@ class Kamal::Cli::Main < Kamal::Cli::Base
       end
     end
 
-    run_hook "post-deploy", secrets: true, runtime: runtime.round
+    run_hook "post-deploy", secrets: true, runtime: runtime.round.to_s
   end
 
-  desc "redeploy", "Deploy app to servers without bootstrapping servers, starting Traefik, pruning, and registry login"
+  desc "redeploy", "Deploy app to servers without bootstrapping servers, starting kamal-proxy, pruning, and registry login"
   option :skip_push, aliases: "-P", type: :boolean, default: false, desc: "Skip image build and push"
   def redeploy
     runtime = print_runtime do
@@ -75,7 +75,7 @@ class Kamal::Cli::Main < Kamal::Cli::Base
       end
     end
 
-    run_hook "post-deploy", secrets: true, runtime: runtime.round
+    run_hook "post-deploy", secrets: true, runtime: runtime.round.to_s
   end
 
   desc "rollback [VERSION]", "Rollback app to VERSION"
@@ -99,12 +99,12 @@ class Kamal::Cli::Main < Kamal::Cli::Base
       end
     end
 
-    run_hook "post-deploy", secrets: true, runtime: runtime.round if rolled_back
+    run_hook "post-deploy", secrets: true, runtime: runtime.round.to_s if rolled_back
   end
 
   desc "details", "Show details about all containers"
   def details
-    invoke "kamal:cli:traefik:details"
+    invoke "kamal:cli:proxy:details"
     invoke "kamal:cli:app:details"
     invoke "kamal:cli:accessory:details", [ "all" ]
   end
@@ -123,7 +123,7 @@ class Kamal::Cli::Main < Kamal::Cli::Base
     end
   end
 
-  desc "docs", "Show Kamal documentation for configuration setting"
+  desc "docs [SECTION]", "Show Kamal configuration documentation"
   def docs(section = nil)
     case section
     when NilClass
@@ -176,19 +176,50 @@ class Kamal::Cli::Main < Kamal::Cli::Base
     end
   end
 
-  desc "remove", "Remove Traefik, app, accessories, and registry session from servers"
+  desc "remove", "Remove kamal-proxy, app, accessories, and registry session from servers"
   option :confirmed, aliases: "-y", type: :boolean, default: false, desc: "Proceed without confirmation question"
   def remove
     confirming "This will remove all containers and images. Are you sure?" do
       with_lock do
-        invoke "kamal:cli:traefik:remove", [], options.without(:confirmed)
         invoke "kamal:cli:app:remove", [], options.without(:confirmed)
+        invoke "kamal:cli:proxy:remove", [], options.without(:confirmed)
         invoke "kamal:cli:accessory:remove", [ "all" ], options
         invoke "kamal:cli:registry:logout", [], options.without(:confirmed).merge(skip_local: true)
       end
     end
   end
 
+  desc "upgrade", "Upgrade from Kamal 1.x to 2.0"
+  option :confirmed, aliases: "-y", type: :boolean, default: false, desc: "Proceed without confirmation question"
+  option :rolling, type: :boolean, default: false, desc: "Upgrade one host at a time"
+  def upgrade
+    confirming "This will replace Traefik with kamal-proxy and restart all accessories" do
+      with_lock do
+        if options[:rolling]
+          (KAMAL.hosts | KAMAL.accessory_hosts).each do |host|
+            KAMAL.with_specific_hosts(host) do
+              say "Upgrading #{host}...", :magenta
+              if KAMAL.hosts.include?(host)
+                invoke "kamal:cli:proxy:upgrade", [], options.merge(confirmed: true, rolling: false)
+                reset_invocation(Kamal::Cli::Proxy)
+              end
+              if KAMAL.accessory_hosts.include?(host)
+                invoke "kamal:cli:accessory:upgrade", [ "all" ], options.merge(confirmed: true, rolling: false)
+                reset_invocation(Kamal::Cli::Accessory)
+              end
+              say "Upgraded #{host}", :magenta
+            end
+          end
+        else
+          say "Upgrading all hosts...", :magenta
+          invoke "kamal:cli:proxy:upgrade", [], options.merge(confirmed: true)
+          invoke "kamal:cli:accessory:upgrade", [ "all" ], options.merge(confirmed: true)
+          say "Upgraded all hosts", :magenta
+        end
+      end
+    end
+  end
+
   desc "version", "Show Kamal version"
   def version
     puts Kamal::VERSION
@@ -206,6 +237,9 @@ class Kamal::Cli::Main < Kamal::Cli::Base
   desc "lock", "Manage the deploy lock"
   subcommand "lock", Kamal::Cli::Lock
 
+  desc "proxy", "Manage kamal-proxy"
+  subcommand "proxy", Kamal::Cli::Proxy
+
   desc "prune", "Prune old application images and containers"
   subcommand "prune", Kamal::Cli::Prune
 
@@ -218,9 +252,6 @@ class Kamal::Cli::Main < Kamal::Cli::Base
   desc "server", "Bootstrap servers with curl and Docker"
   subcommand "server", Kamal::Cli::Server
 
-  desc "traefik", "Manage Traefik load balancer"
-  subcommand "traefik", Kamal::Cli::Traefik
-
   private
     def container_available?(version)
       begin

lib/kamal/cli/proxy.rb

@@ -0,0 +1,215 @@
+class Kamal::Cli::Proxy < Kamal::Cli::Base
+  desc "boot", "Boot proxy on servers"
+  def boot
+    with_lock do
+      on(KAMAL.hosts) do |host|
+        execute *KAMAL.docker.create_network
+      rescue SSHKit::Command::Failed => e
+        raise unless e.message.include?("already exists")
+      end
+
+      on(KAMAL.proxy_hosts) do |host|
+        execute *KAMAL.registry.login
+
+        version = capture_with_info(*KAMAL.proxy.version).strip.presence
+
+        if version && Kamal::Utils.older_version?(version, Kamal::Configuration::PROXY_MINIMUM_VERSION)
+          raise "kamal-proxy version #{version} is too old, please reboot to update to at least #{Kamal::Configuration::PROXY_MINIMUM_VERSION}"
+        end
+        execute *KAMAL.proxy.start_or_run
+      end
+    end
+  end
+
+  desc "reboot", "Reboot proxy on servers (stop container, remove container, start new container)"
+  option :rolling, type: :boolean, default: false, desc: "Reboot proxy on hosts in sequence, rather than in parallel"
+  option :confirmed, aliases: "-y", type: :boolean, default: false, desc: "Proceed without confirmation question"
+  def reboot
+    confirming "This will cause a brief outage on each host. Are you sure?" do
+      with_lock do
+        host_groups = options[:rolling] ? KAMAL.proxy_hosts : [ KAMAL.proxy_hosts ]
+        host_groups.each do |hosts|
+          host_list = Array(hosts).join(",")
+          run_hook "pre-proxy-reboot", hosts: host_list
+          on(hosts) do |host|
+            execute *KAMAL.auditor.record("Rebooted proxy"), verbosity: :debug
+            execute *KAMAL.registry.login
+
+            "Stopping and removing Traefik on #{host}, if running..."
+            execute *KAMAL.proxy.cleanup_traefik
+
+            "Stopping and removing kamal-proxy on #{host}, if running..."
+            execute *KAMAL.proxy.stop, raise_on_non_zero_exit: false
+            execute *KAMAL.proxy.remove_container
+
+            execute *KAMAL.proxy.run
+
+            KAMAL.roles_on(host).select(&:running_proxy?).each do |role|
+              app = KAMAL.app(role: role, host: host)
+
+              version = capture_with_info(*app.current_running_version, raise_on_non_zero_exit: false).strip
+              endpoint = capture_with_info(*app.container_id_for_version(version)).strip
+
+              if endpoint.present?
+                info "Deploying #{endpoint} for role `#{role}` on #{host}..."
+                execute *app.deploy(target: endpoint)
+              end
+            end
+          end
+          run_hook "post-proxy-reboot", hosts: host_list
+        end
+      end
+    end
+  end
+
+  desc "upgrade", "Upgrade to kamal-proxy on servers (stop container, remove container, start new container, reboot app)", hide: true
+  option :rolling, type: :boolean, default: false, desc: "Reboot proxy on hosts in sequence, rather than in parallel"
+  option :confirmed, aliases: "-y", type: :boolean, default: false, desc: "Proceed without confirmation question"
+  def upgrade
+    invoke_options = { "version" => KAMAL.config.latest_tag }.merge(options)
+
+    confirming "This will cause a brief outage on each host. Are you sure?" do
+      host_groups = options[:rolling] ? KAMAL.hosts : [ KAMAL.hosts ]
+      host_groups.each do |hosts|
+        host_list = Array(hosts).join(",")
+        say "Upgrading proxy on #{host_list}...", :magenta
+        run_hook "pre-proxy-reboot", hosts: host_list
+        on(hosts) do |host|
+          execute *KAMAL.auditor.record("Rebooted proxy"), verbosity: :debug
+          execute *KAMAL.registry.login
+
+          "Stopping and removing Traefik on #{host}, if running..."
+          execute *KAMAL.proxy.cleanup_traefik
+
+          "Stopping and removing kamal-proxy on #{host}, if running..."
+          execute *KAMAL.proxy.stop, raise_on_non_zero_exit: false
+          execute *KAMAL.proxy.remove_container
+          execute *KAMAL.proxy.remove_image
+        end
+
+        KAMAL.with_specific_hosts(hosts) do
+          invoke "kamal:cli:proxy:boot", [], invoke_options
+          reset_invocation(Kamal::Cli::Proxy)
+          invoke "kamal:cli:app:boot", [], invoke_options
+          reset_invocation(Kamal::Cli::App)
+          invoke "kamal:cli:prune:all", [], invoke_options
+          reset_invocation(Kamal::Cli::Prune)
+        end
+
+        run_hook "post-proxy-reboot", hosts: host_list
+        say "Upgraded proxy on #{host_list}", :magenta
+      end
+    end
+  end
+
+  desc "start", "Start existing proxy container on servers"
+  def start
+    with_lock do
+      on(KAMAL.proxy_hosts) do |host|
+        execute *KAMAL.auditor.record("Started proxy"), verbosity: :debug
+        execute *KAMAL.proxy.start
+      end
+    end
+  end
+
+  desc "stop", "Stop existing proxy container on servers"
+  def stop
+    with_lock do
+      on(KAMAL.proxy_hosts) do |host|
+        execute *KAMAL.auditor.record("Stopped proxy"), verbosity: :debug
+        execute *KAMAL.proxy.stop, raise_on_non_zero_exit: false
+      end
+    end
+  end
+
+  desc "restart", "Restart existing proxy container on servers"
+  def restart
+    with_lock do
+      stop
+      start
+    end
+  end
+
+  desc "details", "Show details about proxy container from servers"
+  def details
+    on(KAMAL.proxy_hosts) { |host| puts_by_host host, capture_with_info(*KAMAL.proxy.info), type: "Proxy" }
+  end
+
+  desc "logs", "Show log lines from proxy on servers"
+  option :since, aliases: "-s", desc: "Show logs since timestamp (e.g. 2013-01-02T13:23:37Z) or relative (e.g. 42m for 42 minutes)"
+  option :lines, type: :numeric, aliases: "-n", desc: "Number of log lines to pull from each server"
+  option :grep, aliases: "-g", desc: "Show lines with grep match only (use this to fetch specific requests by id)"
+  option :follow, aliases: "-f", desc: "Follow logs on primary server (or specific host set by --hosts)"
+  option :skip_timestamps, type: :boolean, aliases: "-T", desc: "Skip appending timestamps to logging output"
+  def logs
+    grep = options[:grep]
+    timestamps = !options[:skip_timestamps]
+
+    if options[:follow]
+      run_locally do
+        info "Following logs on #{KAMAL.primary_host}..."
+        info KAMAL.proxy.follow_logs(host: KAMAL.primary_host, timestamps: timestamps, grep: grep)
+        exec KAMAL.proxy.follow_logs(host: KAMAL.primary_host, timestamps: timestamps, grep: grep)
+      end
+    else
+      since = options[:since]
+      lines = options[:lines].presence || ((since || grep) ? nil : 100) # Default to 100 lines if since or grep isn't set
+
+      on(KAMAL.proxy_hosts) do |host|
+        puts_by_host host, capture(*KAMAL.proxy.logs(timestamps: timestamps, since: since, lines: lines, grep: grep)), type: "Proxy"
+      end
+    end
+  end
+
+  desc "remove", "Remove proxy container and image from servers"
+  option :force, type: :boolean, default: false, desc: "Force removing proxy when apps are still installed"
+  def remove
+    with_lock do
+      if removal_allowed?(options[:force])
+        stop
+        remove_container
+        remove_image
+      end
+    end
+  end
+
+  desc "remove_container", "Remove proxy container from servers", hide: true
+  def remove_container
+    with_lock do
+      on(KAMAL.proxy_hosts) do
+        execute *KAMAL.auditor.record("Removed proxy container"), verbosity: :debug
+        execute *KAMAL.proxy.remove_container
+      end
+    end
+  end
+
+  desc "remove_image", "Remove proxy image from servers", hide: true
+  def remove_image
+    with_lock do
+      on(KAMAL.proxy_hosts) do
+        execute *KAMAL.auditor.record("Removed proxy image"), verbosity: :debug
+        execute *KAMAL.proxy.remove_image
+      end
+    end
+  end
+
+  private
+    def removal_allowed?(force)
+      on(KAMAL.proxy_hosts) do |host|
+        app_count = capture_with_info(*KAMAL.server.app_directory_count).chomp.to_i
+        raise "The are other applications installed on #{host}" if app_count > 0
+      end
+
+      true
+    rescue SSHKit::Runner::ExecuteError => e
+      raise unless e.message.include?("The are other applications installed on")
+
+      if force
+        say "Forcing, so removing the proxy, even though other apps are installed", :magenta
+      else
+        say "Not removing the proxy, as other apps are installed, ignore this check with kamal proxy remove --force", :magenta
+      end
+
+      force
+    end
+end

lib/kamal/cli/prune.rb

@@ -28,7 +28,6 @@ class Kamal::Cli::Prune < Kamal::Cli::Base
       on(KAMAL.hosts) do
         execute *KAMAL.auditor.record("Pruned containers"), verbosity: :debug
         execute *KAMAL.prune.app_containers(retain: retain)
-        execute *KAMAL.prune.healthcheck_containers
       end
     end
   end

lib/kamal/cli/server.rb

@@ -36,8 +36,6 @@ class Kamal::Cli::Server < Kamal::Cli::Base
             missing << host
           end
         end
-
-        execute(*KAMAL.server.ensure_run_directory)
       end
 
       if missing.any?

lib/kamal/cli/templates/deploy.yml

@@ -57,17 +57,6 @@ builder:
 #     directories:
 #       - data:/data
 
-# Configure custom arguments for Traefik. Be sure to reboot traefik when you modify it.
-# traefik:
-#   args:
-#     accesslog: true
-#     accesslog.format: json
-
-# Configure a custom healthcheck (default is /up on port 3000)
-# healthcheck:
-#   path: /healthz
-#   port: 4000
-
 # Bridge fingerprinted assets, like JS and CSS, between versions to avoid
 # hitting 404 on in-flight requests. Combines all files from new and old
 # version inside the asset_path.

lib/kamal/cli/templates/sample_hooks/post-proxy-reboot.sample

@@ -0,0 +1,3 @@
+#!/bin/sh
+
+echo "Rebooted kamal-proxy on $KAMAL_HOSTS"

lib/kamal/cli/templates/sample_hooks/post-traefik-reboot.sample

@@ -1,3 +0,0 @@
-#!/bin/sh
-
-echo "Rebooted Traefik on $KAMAL_HOSTS"

lib/kamal/cli/traefik.rb

@@ -1,124 +0,0 @@
-class Kamal::Cli::Traefik < Kamal::Cli::Base
-  desc "boot", "Boot Traefik on servers"
-  def boot
-    with_lock do
-      on(KAMAL.traefik_hosts) do
-        execute *KAMAL.registry.login
-        execute *KAMAL.traefik.ensure_env_directory
-        upload! KAMAL.traefik.secrets_io, KAMAL.traefik.secrets_path, mode: "0600"
-        execute *KAMAL.traefik.start_or_run
-      end
-    end
-  end
-
-  desc "reboot", "Reboot Traefik on servers (stop container, remove container, start new container)"
-  option :rolling, type: :boolean, default: false, desc: "Reboot traefik on hosts in sequence, rather than in parallel"
-  option :confirmed, aliases: "-y", type: :boolean, default: false, desc: "Proceed without confirmation question"
-  def reboot
-    confirming "This will cause a brief outage on each host. Are you sure?" do
-      with_lock do
-        host_groups = options[:rolling] ? KAMAL.traefik_hosts : [ KAMAL.traefik_hosts ]
-        host_groups.each do |hosts|
-          host_list = Array(hosts).join(",")
-          run_hook "pre-traefik-reboot", hosts: host_list
-          on(hosts) do
-            execute *KAMAL.auditor.record("Rebooted traefik"), verbosity: :debug
-            execute *KAMAL.registry.login
-            execute *KAMAL.traefik.stop, raise_on_non_zero_exit: false
-            execute *KAMAL.traefik.remove_container
-            execute *KAMAL.traefik.run
-          end
-          run_hook "post-traefik-reboot", hosts: host_list
-        end
-      end
-    end
-  end
-
-  desc "start", "Start existing Traefik container on servers"
-  def start
-    with_lock do
-      on(KAMAL.traefik_hosts) do
-        execute *KAMAL.auditor.record("Started traefik"), verbosity: :debug
-        execute *KAMAL.traefik.start
-      end
-    end
-  end
-
-  desc "stop", "Stop existing Traefik container on servers"
-  def stop
-    with_lock do
-      on(KAMAL.traefik_hosts) do
-        execute *KAMAL.auditor.record("Stopped traefik"), verbosity: :debug
-        execute *KAMAL.traefik.stop, raise_on_non_zero_exit: false
-      end
-    end
-  end
-
-  desc "restart", "Restart existing Traefik container on servers"
-  def restart
-    with_lock do
-      stop
-      start
-    end
-  end
-
-  desc "details", "Show details about Traefik container from servers"
-  def details
-    on(KAMAL.traefik_hosts) { |host| puts_by_host host, capture_with_info(*KAMAL.traefik.info), type: "Traefik" }
-  end
-
-  desc "logs", "Show log lines from Traefik on servers"
-  option :since, aliases: "-s", desc: "Show logs since timestamp (e.g. 2013-01-02T13:23:37Z) or relative (e.g. 42m for 42 minutes)"
-  option :lines, type: :numeric, aliases: "-n", desc: "Number of log lines to pull from each server"
-  option :grep, aliases: "-g", desc: "Show lines with grep match only (use this to fetch specific requests by id)"
-  option :grep_options, aliases: "-o", desc: "Additional options supplied to grep"
-  option :follow, aliases: "-f", desc: "Follow logs on primary server (or specific host set by --hosts)"
-  def logs
-    grep = options[:grep]
-    grep_options = options[:grep_options]
-
-    if options[:follow]
-      run_locally do
-        info "Following logs on #{KAMAL.primary_host}..."
-        info KAMAL.traefik.follow_logs(host: KAMAL.primary_host, grep: grep, grep_options: grep_options)
-        exec KAMAL.traefik.follow_logs(host: KAMAL.primary_host, grep: grep, grep_options: grep_options)
-      end
-    else
-      since = options[:since]
-      lines = options[:lines].presence || ((since || grep) ? nil : 100) # Default to 100 lines if since or grep isn't set
-
-      on(KAMAL.traefik_hosts) do |host|
-        puts_by_host host, capture(*KAMAL.traefik.logs(since: since, lines: lines, grep: grep, grep_options: grep_options)), type: "Traefik"
-      end
-    end
-  end
-
-  desc "remove", "Remove Traefik container and image from servers"
-  def remove
-    with_lock do
-      stop
-      remove_container
-      remove_image
-    end
-  end
-
-  desc "remove_container", "Remove Traefik container from servers", hide: true
-  def remove_container
-    with_lock do
-      on(KAMAL.traefik_hosts) do
-        execute *KAMAL.auditor.record("Removed traefik container"), verbosity: :debug
-        execute *KAMAL.traefik.remove_container
-      end
-    end
-  end
-
-  desc "remove_image", "Remove Traefik image from servers", hide: true
-  def remove_image
-    with_lock do
-      on(KAMAL.traefik_hosts) do
-        execute *KAMAL.auditor.record("Removed traefik image"), verbosity: :debug
-        execute *KAMAL.traefik.remove_image
-      end
-    end
-  end
-end

lib/kamal/commander.rb

@@ -4,7 +4,7 @@ require "active_support/core_ext/object/blank"
 
 class Kamal::Commander
   attr_accessor :verbosity, :holding_lock, :connected
-  delegate :hosts, :roles, :primary_host, :primary_role, :roles_on, :traefik_hosts, :accessory_hosts, to: :specifics
+  delegate :hosts, :roles, :primary_host, :primary_role, :roles_on, :proxy_hosts, :accessory_hosts, to: :specifics
 
   def initialize
     self.verbosity = :info
@@ -65,6 +65,13 @@ class Kamal::Commander
     end
   end
 
+  def with_specific_hosts(hosts)
+    original_hosts, self.specific_hosts = specific_hosts, hosts
+    yield
+  ensure
+    self.specific_hosts = original_hosts
+  end
+
   def accessory_names
     config.accessories&.collect(&:name) || []
   end
@@ -94,10 +101,6 @@ class Kamal::Commander
     @docker ||= Kamal::Commands::Docker.new(config)
   end
 
-  def healthcheck
-    @healthcheck ||= Kamal::Commands::Healthcheck.new(config)
-  end
-
   def hook
     @hook ||= Kamal::Commands::Hook.new(config)
   end
@@ -106,6 +109,10 @@ class Kamal::Commander
     @lock ||= Kamal::Commands::Lock.new(config)
   end
 
+  def proxy
+    @proxy ||= Kamal::Commands::Proxy.new(config)
+  end
+
   def prune
     @prune ||= Kamal::Commands::Prune.new(config)
   end
@@ -118,10 +125,6 @@ class Kamal::Commander
     @server ||= Kamal::Commands::Server.new(config)
   end
 
-  def traefik
-    @traefik ||= Kamal::Commands::Traefik.new(config)
-  end
-
   def alias(name)
     config.aliases[name]
   end

lib/kamal/commander/specifics.rb

@@ -18,12 +18,12 @@ class Kamal::Commander::Specifics
     roles.select { |role| role.hosts.include?(host.to_s) }
   end
 
-  def traefik_hosts
-    config.traefik_hosts & specified_hosts
+  def proxy_hosts
+    config.proxy_hosts & specified_hosts
   end
 
   def accessory_hosts
-    specific_hosts || config.accessories.flat_map(&:hosts)
+    config.accessories.flat_map(&:hosts) & specified_hosts
   end
 
   private

lib/kamal/commands/accessory.rb

@@ -15,6 +15,7 @@ class Kamal::Commands::Accessory < Kamal::Commands::Base
       "--name", service_name,
       "--detach",
       "--restart", "unless-stopped",
+      "--network", "kamal",
       *config.logging_args,
       *publish_args,
       *env_args,
@@ -38,16 +39,16 @@ class Kamal::Commands::Accessory < Kamal::Commands::Base
   end
 
 
-  def logs(since: nil, lines: nil, grep: nil, grep_options: nil)
+  def logs(timestamps: true, since: nil, lines: nil, grep: nil, grep_options: nil)
     pipe \
-      docker(:logs, service_name, (" --since #{since}" if since), (" --tail #{lines}" if lines), "--timestamps", "2>&1"),
+      docker(:logs, service_name, (" --since #{since}" if since), (" --tail #{lines}" if lines), ("--timestamps" if timestamps), "2>&1"),
       ("grep '#{grep}'#{" #{grep_options}" if grep_options}" if grep)
   end
 
-  def follow_logs(grep: nil, grep_options: nil)
+  def follow_logs(timestamps: true, grep: nil, grep_options: nil)
     run_over_ssh \
       pipe \
-        docker(:logs, service_name, "--timestamps", "--tail", "10", "--follow", "2>&1"),
+        docker(:logs, service_name, ("--timestamps" if timestamps), "--tail", "10", "--follow", "2>&1"),
         (%(grep "#{grep}"#{" #{grep_options}" if grep_options}) if grep)
   end
 
@@ -63,6 +64,7 @@ class Kamal::Commands::Accessory < Kamal::Commands::Base
     docker :run,
       ("-it" if interactive),
       "--rm",
+      "--network", "kamal",
       *env_args,
       *volume_args,
       image,

lib/kamal/commands/app.rb

@@ -1,10 +1,12 @@
 class Kamal::Commands::App < Kamal::Commands::Base
-  include Assets, Containers, Cord, Execution, Images, Logging
+  include Assets, Containers, Execution, Images, Logging, Proxy
 
   ACTIVE_DOCKER_STATUSES = [ :running, :restarting ]
 
   attr_reader :role, :host
 
+  delegate :container_name, to: :role
+
   def initialize(config, role: nil, host: nil)
     super(config)
     @role = role
@@ -16,11 +18,11 @@ class Kamal::Commands::App < Kamal::Commands::Base
       "--detach",
       "--restart unless-stopped",
       "--name", container_name,
+      "--network", "kamal",
       *([ "--hostname", hostname ] if hostname),
       "-e", "KAMAL_CONTAINER_NAME=\"#{container_name}\"",
       "-e", "KAMAL_VERSION=\"#{config.version}\"",
       *role.env_args(host),
-      *role.health_check_args,
       *role.logging_args,
       *config.volume_args,
       *role.asset_volume_args,
@@ -41,7 +43,7 @@ class Kamal::Commands::App < Kamal::Commands::Base
   def stop(version: nil)
     pipe \
       version ? container_id_for_version(version) : current_running_container_id,
-      xargs(config.stop_wait_time ? docker(:stop, "-t", config.stop_wait_time) : docker(:stop))
+      xargs(docker(:stop, *role.stop_args))
   end
 
   def info
@@ -74,10 +76,6 @@ class Kamal::Commands::App < Kamal::Commands::Base
   end
 
   private
-    def container_name(version = nil)
-      [ role.container_prefix, version || config.version ].compact.join("-")
-    end
-
     def latest_image_id
       docker :image, :ls, *argumentize("--filter", "reference=#{config.latest_image}"), "--format", "'{{.ID}}'"
     end

lib/kamal/commands/app/assets.rb

@@ -3,18 +3,18 @@ module Kamal::Commands::App::Assets
     asset_container = "#{role.container_prefix}-assets"
 
     combine \
-      make_directory(role.asset_extracted_path),
+      make_directory(role.asset_extracted_directory),
       [ *docker(:stop, "-t 1", asset_container, "2> /dev/null"), "|| true" ],
       docker(:run, "--name", asset_container, "--detach", "--rm", "--entrypoint", "sleep", config.absolute_image, "1000000"),
-      docker(:cp, "-L", "#{asset_container}:#{role.asset_path}/.", role.asset_extracted_path),
+      docker(:cp, "-L", "#{asset_container}:#{role.asset_path}/.", role.asset_extracted_directory),
       docker(:stop, "-t 1", asset_container),
       by: "&&"
   end
 
   def sync_asset_volumes(old_version: nil)
-    new_extracted_path, new_volume_path = role.asset_extracted_path(config.version), role.asset_volume.host_path
+    new_extracted_path, new_volume_path = role.asset_extracted_directory(config.version), role.asset_volume.host_path
     if old_version.present?
-      old_extracted_path, old_volume_path = role.asset_extracted_path(old_version), role.asset_volume(old_version).host_path
+      old_extracted_path, old_volume_path = role.asset_extracted_directory(old_version), role.asset_volume(old_version).host_path
     end
 
     commands = [ make_directory(new_volume_path), copy_contents(new_extracted_path, new_volume_path) ]
@@ -29,8 +29,8 @@ module Kamal::Commands::App::Assets
 
   def clean_up_assets
     chain \
-      find_and_remove_older_siblings(role.asset_extracted_path),
-      find_and_remove_older_siblings(role.asset_volume_path)
+      find_and_remove_older_siblings(role.asset_extracted_directory),
+      find_and_remove_older_siblings(role.asset_volume_directory)
   end
 
   private
@@ -39,7 +39,7 @@ module Kamal::Commands::App::Assets
         :find,
         Pathname.new(path).dirname.to_s,
         "-maxdepth 1",
-        "-name", "'#{role.container_prefix}-*'",
+        "-name", "'#{role.name}-*'",
         "!", "-name", Pathname.new(path).basename.to_s,
         "-exec rm -rf \"{}\" +"
       ]

lib/kamal/commands/app/cord.rb

@@ -1,22 +0,0 @@
-module Kamal::Commands::App::Cord
-  def cord(version:)
-    pipe \
-      docker(:inspect, "-f '{{ range .Mounts }}{{printf \"%s %s\\n\" .Source .Destination}}{{ end }}'", container_name(version)),
-      [ :awk, "'$2 == \"#{role.cord_volume.container_path}\" {print $1}'" ]
-  end
-
-  def tie_cord(cord)
-    create_empty_file(cord)
-  end
-
-  def cut_cord(cord)
-    remove_directory(cord)
-  end
-
-  private
-    def create_empty_file(file)
-      chain \
-        make_directory_for(file),
-        [ :touch, file ]
-    end
-end

lib/kamal/commands/app/execution.rb

@@ -11,6 +11,7 @@ module Kamal::Commands::App::Execution
     docker :run,
       ("-it" if interactive),
       "--rm",
+      "--network", "kamal",
       *role&.env_args(host),
       *argumentize("--env", env),
       *config.volume_args,

lib/kamal/commands/app/logging.rb

@@ -1,16 +1,16 @@
 module Kamal::Commands::App::Logging
-  def logs(version: nil, since: nil, lines: nil, grep: nil, grep_options: nil)
+  def logs(version: nil, timestamps: true, since: nil, lines: nil, grep: nil, grep_options: nil)
     pipe \
       version ? container_id_for_version(version) : current_running_container_id,
-      "xargs docker logs#{" --since #{since}" if since}#{" --tail #{lines}" if lines} 2>&1",
+      "xargs docker logs#{" --timestamps" if timestamps}#{" --since #{since}" if since}#{" --tail #{lines}" if lines} 2>&1",
       ("grep '#{grep}'#{" #{grep_options}" if grep_options}" if grep)
   end
 
-  def follow_logs(host:, lines: nil, grep: nil, grep_options: nil)
+  def follow_logs(host:, timestamps: true, lines: nil, grep: nil, grep_options: nil)
     run_over_ssh \
       pipe(
         current_running_container_id,
-        "xargs docker logs --timestamps#{" --tail #{lines}" if lines} --follow 2>&1",
+        "xargs docker logs#{" --timestamps" if timestamps}#{" --tail #{lines}" if lines} --follow 2>&1",
         (%(grep "#{grep}"#{" #{grep_options}" if grep_options}) if grep)
       ),
       host: host

lib/kamal/commands/app/proxy.rb

@@ -0,0 +1,16 @@
+module Kamal::Commands::App::Proxy
+  delegate :proxy_container_name, to: :config
+
+  def deploy(target:)
+    proxy_exec :deploy, role.container_prefix, *role.proxy.deploy_command_args(target: target)
+  end
+
+  def remove(target:)
+    proxy_exec :remove, role.container_prefix, *role.proxy.remove_command_args(target: target)
+  end
+
+  private
+    def proxy_exec(*command)
+      docker :exec, proxy_container_name, "kamal-proxy", *command
+    end
+end

lib/kamal/commands/docker.rb

@@ -19,6 +19,10 @@ class Kamal::Commands::Docker < Kamal::Commands::Base
     [ '[ "${EUID:-$(id -u)}" -eq 0 ] || command -v sudo >/dev/null || command -v su >/dev/null' ]
   end
 
+  def create_network
+    docker :network, :create, :kamal
+  end
+
   private
     def get_docker
       shell \

lib/kamal/commands/hook.rb

@@ -1,9 +1,12 @@
 class Kamal::Commands::Hook < Kamal::Commands::Base
-  def run(hook, secrets: false, **details)
-    env = tags(**details).env
-    env.merge!(config.secrets.to_h) if secrets
+  def run(hook)
+    [ hook_file(hook) ]
+  end
 
-    [ hook_file(hook), env: env ]
+  def env(secrets: false, **details)
+    tags(**details).env.tap do |env|
+      env.merge!(config.secrets.to_h) if secrets
+    end
   end
 
   def hook_exists?(hook)

lib/kamal/commands/lock.rb

@@ -44,14 +44,10 @@ class Kamal::Commands::Lock < Kamal::Commands::Base
         "/dev/null"
     end
 
-    def locks_dir
-      File.join(config.run_directory, "locks")
-    end
-
     def lock_dir
-      dir_name = [ config.service, config.destination ].compact.join("-")
+      dir_name = [ "lock", config.service, config.destination ].compact.join("-")
 
-      File.join(locks_dir, dir_name)
+      File.join(config.run_directory, dir_name)
     end
 
     def lock_details_file

lib/kamal/commands/proxy.rb

@@ -0,0 +1,72 @@
+class Kamal::Commands::Proxy < Kamal::Commands::Base
+  delegate :argumentize, :optionize, to: Kamal::Utils
+
+  def run
+    docker :run,
+      "--name", container_name,
+      "--network", "kamal",
+      "--detach",
+      "--restart", "unless-stopped",
+      *config.proxy_publish_args,
+      "--volume", "kamal-proxy-config:/home/kamal-proxy/.config/kamal-proxy",
+      *config.logging_args,
+      config.proxy_image
+  end
+
+  def start
+    docker :container, :start, container_name
+  end
+
+  def stop(name: container_name)
+    docker :container, :stop, name
+  end
+
+  def start_or_run
+    combine start, run, by: "||"
+  end
+
+  def info
+    docker :ps, "--filter", "name=^#{container_name}$"
+  end
+
+  def version
+    pipe \
+      docker(:inspect, container_name, "--format '{{.Config.Image}}'"),
+      [ :cut, "-d:", "-f2" ]
+  end
+
+  def logs(timestamps: true, since: nil, lines: nil, grep: nil, grep_options: nil)
+    pipe \
+      docker(:logs, container_name, ("--since #{since}" if since), ("--tail #{lines}" if lines), ("--timestamps" if timestamps), "2>&1"),
+      ("grep '#{grep}'#{" #{grep_options}" if grep_options}" if grep)
+  end
+
+  def follow_logs(host:, timestamps: true, grep: nil, grep_options: nil)
+    run_over_ssh pipe(
+      docker(:logs, container_name, ("--timestamps" if timestamps), "--tail", "10", "--follow", "2>&1"),
+      (%(grep "#{grep}"#{" #{grep_options}" if grep_options}) if grep)
+    ).join(" "), host: host
+  end
+
+  def remove_container
+    docker :container, :prune, "--force", "--filter", "label=org.opencontainers.image.title=kamal-proxy"
+  end
+
+  def remove_image
+    docker :image, :prune, "--all", "--force", "--filter", "label=org.opencontainers.image.title=kamal-proxy"
+  end
+
+  def cleanup_traefik
+    chain \
+      docker(:container, :stop, "traefik"),
+      combine(
+        docker(:container, :prune, "--force", "--filter", "label=org.opencontainers.image.title=Traefik"),
+        docker(:image, :prune, "--all", "--force", "--filter", "label=org.opencontainers.image.title=Traefik")
+      )
+  end
+
+  private
+    def container_name
+      config.proxy_container_name
+    end
+end

lib/kamal/commands/prune.rb

@@ -20,10 +20,6 @@ class Kamal::Commands::Prune < Kamal::Commands::Base
       "while read container_id; do docker rm $container_id; done"
   end
 
-  def healthcheck_containers
-    docker :container, :prune, "--force", *healthcheck_service_filter
-  end
-
   private
     def stopped_containers_filters
       [ "created", "exited", "dead" ].flat_map { |status| [ "--filter", "status=#{status}" ] }
@@ -39,8 +35,4 @@ class Kamal::Commands::Prune < Kamal::Commands::Base
     def service_filter
       [ "--filter", "label=service=#{config.service}" ]
     end
-
-    def healthcheck_service_filter
-      [ "--filter", "label=service=#{config.healthcheck_service}" ]
-    end
 end

lib/kamal/commands/server.rb

@@ -1,5 +1,15 @@
 class Kamal::Commands::Server < Kamal::Commands::Base
   def ensure_run_directory
-    [ :mkdir, "-p", config.run_directory ]
+    make_directory config.run_directory
+  end
+
+  def remove_app_directory
+    remove_directory config.app_directory
+  end
+
+  def app_directory_count
+    pipe \
+      [ :ls, config.apps_directory ],
+      [ :wc, "-l" ]
   end
 end

lib/kamal/commands/traefik.rb

@@ -1,77 +0,0 @@
-class Kamal::Commands::Traefik < Kamal::Commands::Base
-  delegate :argumentize, :optionize, to: Kamal::Utils
-  delegate :port, :publish?, :labels, :env, :image, :options, :args, :env_args, :secrets_io, :env_directory, :secrets_path, to: :"config.traefik"
-
-  def run
-    docker :run, "--name traefik",
-      "--detach",
-      "--restart", "unless-stopped",
-      *publish_args,
-      "--volume", "/var/run/docker.sock:/var/run/docker.sock",
-      *env_args,
-      *config.logging_args,
-      *label_args,
-      *docker_options_args,
-      image,
-      "--providers.docker",
-      *cmd_option_args
-  end
-
-  def start
-    docker :container, :start, "traefik"
-  end
-
-  def stop
-    docker :container, :stop, "traefik"
-  end
-
-  def start_or_run
-    any start, run
-  end
-
-  def info
-    docker :ps, "--filter", "name=^traefik$"
-  end
-
-  def logs(since: nil, lines: nil, grep: nil, grep_options: nil)
-    pipe \
-      docker(:logs, "traefik", (" --since #{since}" if since), (" --tail #{lines}" if lines), "--timestamps", "2>&1"),
-      ("grep '#{grep}'#{" #{grep_options}" if grep_options}" if grep)
-  end
-
-  def follow_logs(host:, grep: nil, grep_options: nil)
-    run_over_ssh pipe(
-      docker(:logs, "traefik", "--timestamps", "--tail", "10", "--follow", "2>&1"),
-      (%(grep "#{grep}"#{" #{grep_options}" if grep_options}) if grep)
-    ).join(" "), host: host
-  end
-
-  def remove_container
-    docker :container, :prune, "--force", "--filter", "label=org.opencontainers.image.title=Traefik"
-  end
-
-  def remove_image
-    docker :image, :prune, "--all", "--force", "--filter", "label=org.opencontainers.image.title=Traefik"
-  end
-
-  def ensure_env_directory
-    make_directory env_directory
-  end
-
-  private
-    def publish_args
-      argumentize "--publish", port if publish?
-    end
-
-    def label_args
-      argumentize "--label", labels
-    end
-
-    def docker_options_args
-      optionize(options)
-    end
-
-    def cmd_option_args
-      optionize args, with: "="
-    end
-end

lib/kamal/configuration.rb

@@ -6,14 +6,18 @@ require "erb"
 require "net/ssh/proxy/jump"
 
 class Kamal::Configuration
-  delegate :service, :image, :labels, :stop_wait_time, :hooks_path, to: :raw_config, allow_nil: true
+  delegate :service, :image, :labels, :hooks_path, to: :raw_config, allow_nil: true
   delegate :argumentize, :optionize, to: Kamal::Utils
 
   attr_reader :destination, :raw_config, :secrets
-  attr_reader :accessories, :aliases, :boot, :builder, :env, :healthcheck, :logging, :traefik, :servers, :ssh, :sshkit, :registry
+  attr_reader :accessories, :aliases, :boot, :builder, :env, :logging, :proxy, :servers, :ssh, :sshkit, :registry
 
   include Validation
 
+  PROXY_MINIMUM_VERSION = "v0.4.0"
+  PROXY_HTTP_PORT = 80
+  PROXY_HTTPS_PORT = 443
+
   class << self
     def create_from(config_file:, destination: nil, version: nil)
       raw_config = load_config_files(config_file, *destination_config_file(config_file, destination))
@@ -48,6 +52,8 @@ class Kamal::Configuration
 
     validate! raw_config, example: validation_yml.symbolize_keys, context: "", with: Kamal::Configuration::Validator::Configuration
 
+    @secrets = Kamal::Secrets.new(destination: destination)
+
     # Eager load config to validate it, these are first as they have dependencies later on
     @servers = Servers.new(config: self)
     @registry = Registry.new(config: self)
@@ -58,19 +64,19 @@ class Kamal::Configuration
     @builder = Builder.new(config: self)
     @env = Env.new(config: @raw_config.env || {}, secrets: secrets)
 
-    @healthcheck = Healthcheck.new(healthcheck_config: @raw_config.healthcheck)
     @logging = Logging.new(logging_config: @raw_config.logging)
-    @traefik = Traefik.new(config: self)
+    @proxy = Proxy.new(config: self, proxy_config: @raw_config.proxy || {})
     @ssh = Ssh.new(config: self)
     @sshkit = Sshkit.new(config: self)
 
-    @secrets = Kamal::Secrets.new(destination: destination)
-
     ensure_destination_if_required
     ensure_required_keys_present
     ensure_valid_kamal_version
     ensure_retain_containers_valid
     ensure_valid_service_name
+    ensure_no_traefik_reboot_hooks
+    ensure_one_host_for_ssl_roles
+    ensure_unique_hosts_for_ssl_roles
   end
 
 
@@ -131,16 +137,16 @@ class Kamal::Configuration
     raw_config.allow_empty_roles
   end
 
-  def traefik_roles
-    roles.select(&:running_traefik?)
+  def proxy_roles
+    roles.select(&:running_proxy?)
   end
 
-  def traefik_role_names
-    traefik_roles.flat_map(&:name)
+  def proxy_role_names
+    proxy_roles.flat_map(&:name)
   end
 
-  def traefik_hosts
-    traefik_roles.flat_map(&:hosts).uniq
+  def proxy_hosts
+    proxy_roles.flat_map(&:hosts).uniq
   end
 
   def repository
@@ -185,31 +191,40 @@ class Kamal::Configuration
   end
 
 
-  def healthcheck_service
-    [ "healthcheck", service, destination ].compact.join("-")
-  end
-
   def readiness_delay
     raw_config.readiness_delay || 7
   end
 
-  def run_id
-    @run_id ||= SecureRandom.hex(16)
+  def deploy_timeout
+    raw_config.deploy_timeout || 30
+  end
+
+  def drain_timeout
+    raw_config.drain_timeout || 30
   end
 
 
   def run_directory
-    raw_config.run_directory || ".kamal"
+    ".kamal"
   end
 
-  def run_directory_as_docker_volume
-    if Pathname.new(run_directory).absolute?
-      run_directory
-    else
-      File.join "$(pwd)", run_directory
-    end
+  def apps_directory
+    File.join run_directory, "apps"
   end
 
+  def app_directory
+    File.join apps_directory, [ service, destination ].compact.join("-")
+  end
+
+  def env_directory
+    File.join app_directory, "env"
+  end
+
+  def assets_directory
+    File.join app_directory, "assets"
+  end
+
+
   def hooks_path
     raw_config.hooks_path || ".kamal/hooks"
   end
@@ -219,10 +234,6 @@ class Kamal::Configuration
   end
 
 
-  def env_directory
-    File.join(run_directory, "env")
-  end
-
   def env_tags
     @env_tags ||= if (tags = raw_config.env["tags"])
       tags.collect { |name, config| Env::Tag.new(name, config: config, secrets: secrets) }
@@ -235,6 +246,18 @@ class Kamal::Configuration
     env_tags.detect { |t| t.name == name.to_s }
   end
 
+  def proxy_publish_args
+    argumentize "--publish", [ "#{PROXY_HTTP_PORT}:#{PROXY_HTTP_PORT}", "#{PROXY_HTTPS_PORT}:#{PROXY_HTTPS_PORT}" ]
+  end
+
+  def proxy_image
+    "basecamp/kamal-proxy:#{PROXY_MINIMUM_VERSION}"
+  end
+
+  def proxy_container_name
+    "kamal-proxy"
+  end
+
 
   def to_h
     {
@@ -250,15 +273,10 @@ class Kamal::Configuration
       sshkit: sshkit.to_h,
       builder: builder.to_h,
       accessories: raw_config.accessories,
-      logging: logging_args,
-      healthcheck: healthcheck.to_h
+      logging: logging_args
     }.compact
   end
 
-  def secrets
-    @secrets ||= Kamal::Secrets.new(destination: destination)
-  end
-
   private
     # Will raise ArgumentError if any required config keys are missing
     def ensure_destination_if_required
@@ -313,6 +331,30 @@ class Kamal::Configuration
       true
     end
 
+    def ensure_no_traefik_reboot_hooks
+      hooks = %w[ pre-traefik-reboot post-traefik-reboot ].select { |hook_file| File.exist?(File.join(hooks_path, hook_file)) }
+
+      if hooks.any?
+        raise Kamal::ConfigurationError, "Found #{hooks.join(", ")}, these should be renamed to (pre|post)-proxy-reboot"
+      end
+
+      true
+    end
+
+    def ensure_one_host_for_ssl_roles
+      roles.each(&:ensure_one_host_for_ssl)
+
+      true
+    end
+
+    def ensure_unique_hosts_for_ssl_roles
+      hosts = roles.select(&:ssl?).map { |role| role.proxy.host }
+      duplicates = hosts.tally.filter_map { |host, count| host if count > 1 }
+
+      raise Kamal::ConfigurationError, "Different roles can't share the same host for SSL: #{duplicates.join(", ")}" if duplicates.any?
+
+      true
+    end
 
     def role_names
       raw_config.servers.is_a?(Array) ? [ "web" ] : raw_config.servers.keys.sort

lib/kamal/configuration/accessory.rb

@@ -63,7 +63,7 @@ class Kamal::Configuration::Accessory
   end
 
   def secrets_path
-    File.join(config.env_directory, "accessories", "#{service_name}.env")
+    File.join(config.env_directory, "accessories", "#{name}.env")
   end
 
   def files

lib/kamal/configuration/docs/builder.yml

@@ -2,35 +2,31 @@
 #
 # The builder configuration controls how the application is built with `docker build`
 #
-# If no configuration is specified, Kamal will:
-# 1. Create a buildx context called `kamal-local-docker-container`, using the docker-container driver
-# 2. Use `docker build` to build a multiarch image for linux/amd64,linux/arm64 with that context
-#
 # See https://kamal-deploy.org/docs/configuration/builder-examples/ for more information
 
 # Builder options
 #
 # Options go under the builder key in the root configuration.
 builder:
-  # Driver
-  #
-  # The build driver to use, defaults to `docker-container`
-  driver: docker
 
   # Arch
   #
-  # The architectures to build for, defaults to `[ amd64, arm64 ]`
-  # Unless you are using the docker driver, when it defaults to the local architecture
-  # You can set an array or just a single value
+  # The architectures to build for - you can set an array or just a single value.
+  #
+  # Allowed values are `amd64` and `arm64`
   arch:
     - amd64
 
-  # Remote configuration
+  # Remote
   #
-  # If you have a remote builder, you can configure it here
+  # The connection string for a remote builder. If supplied Kamal will use this
+  # for builds that do not match the local architecture of the deployment host.
   remote: ssh://docker@docker-builder
 
-  # Whether to allow local builds
+  # Local
+  #
+  # If set to false, Kamal will always use the remote builder even when building
+  # the local architecture.
   #
   # Defaults to true
   local: true
@@ -87,7 +83,7 @@ builder:
 
   # Build secrets
   #
-  # Values are read from the environment.
+  # Values are read from .kamal/secrets.
   #
   secrets:
     - SECRET1
@@ -112,3 +108,8 @@ builder:
   #
   # SSH agent socket or keys to expose to the build
   ssh: default=$SSH_AUTH_SOCK
+
+  # Driver
+  #
+  # The build driver to use, defaults to `docker-container`
+  driver: docker

lib/kamal/configuration/docs/configuration.yml

@@ -36,6 +36,8 @@ image: my-image
 labels:
   my-label: my-value
 
+# Volumes
+#
 # Additional volumes to mount into the container
 volumes:
   - /path/on/host:/path/in/container:ro
@@ -58,7 +60,7 @@ servers:
 env:
   ...
 
-# Asset Bridging
+# Asset Path
 #
 # Used for asset bridging across deployments, default to `nil`
 #
@@ -70,10 +72,12 @@ env:
 # volume containing both sets of files.
 # This requires that file names change when the contents change
 # (e.g. by including a hash of the contents in the name).
-
+#
 # To configure this, set the path to the assets:
 asset_path: /path/to/assets
 
+# Hooks path
+#
 # Path to hooks, defaults to `.kamal/hooks`
 # See https://kamal-deploy.org/docs/hooks for more information
 hooks_path: /user_home/kamal/hooks
@@ -83,7 +87,7 @@ hooks_path: /user_home/kamal/hooks
 # Whether deployments require a destination to be specified, defaults to `false`
 require_destination: true
 
-# The primary role
+# Primary role
 #
 # This defaults to `web`, but if you have no web role, you can change this
 primary_role: workers
@@ -93,11 +97,6 @@ primary_role: workers
 # Whether roles with no servers are allowed. Defaults to `false`.
 allow_empty_roles: false
 
-# Stop wait time
-#
-# How long we wait for a container to stop before killing it, defaults to 30 seconds
-stop_wait_time: 60
-
 # Retain containers
 #
 # How many old containers and images we retain, defaults to 5
@@ -111,9 +110,20 @@ minimum_version: 1.3.0
 # Readiness delay
 #
 # Seconds to wait for a container to boot after is running, default 7
-# This only applies to containers that do not specify a healthcheck
+#
+# This only applies to containers that do not run a proxy or specify a healthcheck
 readiness_delay: 4
 
+# Deploy timeout
+#
+# How long to wait for a container to become ready, default 30
+deploy_timeout: 10
+
+# Drain timeout
+#
+# How long to wait for a containers to drain, default 30
+drain_timeout: 10
+
 # Run directory
 #
 # Directory to store kamal runtime files in on the host, default `.kamal`
@@ -137,10 +147,10 @@ builder:
 accessories:
   ...
 
-# Traefik
+# Proxy
 #
-# The Traefik proxy is used for zero-downtime deployments, see kamal docs traefik
-traefik:
+# Configuration for kamal-proxy, see kamal docs proxy
+proxy:
   ...
 
 # SSHKit
@@ -155,12 +165,6 @@ sshkit:
 boot:
   ...
 
-# Healthcheck
-#
-# Configuring healthcheck commands, intervals and timeouts, see kamal docs healthcheck
-healthcheck:
-  ...
-
 # Logging
 #
 # Docker logging configuration, see kamal docs logging

lib/kamal/configuration/docs/env.yml

@@ -1,7 +1,7 @@
 # Environment variables
 #
 # Environment variables can be set directly in the Kamal configuration or
-# loaded from a .env file, for secrets that should not be checked into Git.
+# read from .kamal/secrets.
 
 # Reading environment variables from the configuration
 #
@@ -12,26 +12,38 @@ env:
   DATABASE_HOST: mysql-db1
   DATABASE_PORT: 3306
 
-# Using .env file to load required environment variables
+# Secrets
 #
-# Kamal uses dotenv to automatically load environment variables set in the .env file present
-# in the application root.
+# Kamal uses dotenv to automatically load environment variables set in the `.kamal/secrets` file.
+#
+# If you are using destinations, secrets will instead be read from `.kamal/secrets-<DESTINATION>` if
+# it exists.
+#
+# Common secrets across all destinations can be set in `.kamal/secrets-common`.
+#
+# This file can be used to set variables like `KAMAL_REGISTRY_PASSWORD` or database passwords.
+# You can use variable or command substitution in the secrets file.
 #
-# This file can be used to set variables like KAMAL_REGISTRY_PASSWORD or database passwords.
-# But for this reason you must ensure that .env files are not checked into Git or included
-# in your Dockerfile! The format is just key-value like:
 # ```
-# KAMAL_REGISTRY_PASSWORD=pw
-# DB_PASSWORD=secret123
+# KAMAL_REGISTRY_PASSWORD=$KAMAL_REGISTRY_PASSWORD
+# RAILS_MASTER_KEY=$(cat config/master.key)
 # ```
-# See https://kamal-deploy.org/docs/commands/envify/ for how to use generated .env files.
+#
+# You can also use [secret helpers](../commands/secrets) for some common password managers.
+# ```
+# SECRETS=$(kamal secrets fetch ...)
+#
+# REGISTRY_PASSWORD=$(kamal secrets extract REGISTRY_PASSWORD $SECRETS)
+# DB_PASSWORD=$(kamal secrets extract DB_PASSWORD $SECRETS)
+# ```
+#
+# If you store secrets directly in .kamal/secrets, ensure that it is not checked into version control.
 #
 # To pass the secrets you should list them under the `secret` key. When you do this the
 # other variables need to be moved under the `clear` key.
 #
 # Unlike clear values, secrets are not passed directly to the container,
 # but are stored in an env file on the host
-# The file is not updated when deploying, only when running `kamal envify` or `kamal env push`.
 env:
   clear:
     DB_USER: app

lib/kamal/configuration/docs/healthcheck.yml

@@ -1,59 +0,0 @@
-# Healthcheck configuration
-#
-# On roles that are running Traefik, Kamal will supply a default healthcheck to `docker run`.
-# For other roles, by default no healthcheck is supplied.
-#
-# If no healthcheck is supplied and the image does not define one, then we wait for the container
-# to reach a running state and then pause for the readiness delay.
-#
-# The default healthcheck is `curl -f http://localhost:<port>/<path>`, so it assumes that `curl`
-# is available within the container.
-
-# Healthcheck options
-#
-# These go under the `healthcheck` key in the root or role configuration.
-healthcheck:
-
-  # Command
-  #
-  # The command to run, defaults to `curl -f http://localhost:<port>/<path>` on roles running Traefik
-  cmd: "curl -f http://localhost"
-
-  # Interval
-  #
-  # The Docker healthcheck interval, defaults to `1s`
-  interval: 10s
-
-  # Max attempts
-  #
-  # The maximum number of times we poll the container to see if it is healthy, defaults to `7`
-  # Each check is separated by an increasing interval starting with 1 second.
-  max_attempts: 3
-
-  # Port
-  #
-  # The port to use in the healthcheck, defaults to `3000`
-  port: "80"
-
-  # Path
-  #
-  # The path to use in the healthcheck, defaults to `/up`
-  path: /health
-
-  # Cords for zero-downtime deployments
-  #
-  # The cord file is used for zero-downtime deployments. The healthcheck is augmented with a check
-  # for the existance of the file. This allows us to delete the file and force the container to
-  # become unhealthy, causing Traefik to stop routing traffic to it.
-  #
-  # Kamal mounts a volume at this location and creates the file before starting the container.
-  # You can set the value to `false` to disable the cord file, but this loses the zero-downtime
-  # guarantee.
-  #
-  # The default value is `/tmp/kamal-cord`
-  cord: /cord
-
-  # Log lines
-  #
-  # Number of lines to log from the container when the healthcheck fails, defaults to `50`
-  log_lines: 100

lib/kamal/configuration/docs/proxy.yml

@@ -0,0 +1,100 @@
+# Proxy
+#
+# Kamal uses [kamal-proxy](https://github.com/basecamp/kamal-proxy) to provide
+# gapless deployments. It runs on ports 80 and 443 and forwards requests to the
+# application container.
+#
+# The proxy is configured in the root configuration under `proxy`. These are
+# options that are set when deploying the application, not when booting the proxy
+#
+# They are application specific, so are not shared when multiple applications
+# run on the same proxy.
+#
+# The proxy is enabled by default on the primary role, but can be disabled by
+# setting `proxy: false`.
+#
+# It is disabled by default on all other roles, but can be enabled by setting
+# `proxy: true`, or providing a proxy configuration.
+proxy:
+
+  # Host
+  #
+  # The hosts that will be used to serve the app. The proxy will only route requests
+  # to this host to your app.
+  #
+  # If no hosts are set, then all requests will be forwarded, except for matching
+  # requests for other apps deployed on that server that do have a host set.
+  host: foo.example.com
+
+  # App port
+  #
+  # The port the application container is exposed on
+  #
+  # Defaults to 80
+  app_port: 3000
+
+  # SSL
+  #
+  # kamal-proxy can provide automatic HTTPS for your application via Let's Encrypt.
+  #
+  # This requires that we are deploying to a one server and the host option is set.
+  # The host value must point to the server we are deploying to and port 443 must be
+  # open for the Let's Encrypt challenge to succeed.
+  #
+  # Defaults to false
+  ssl: true
+
+  # Response timeout
+  #
+  # How long to wait for requests to complete before timing out, defaults to 30 seconds
+  response_timeout: 10
+
+  # Healthcheck
+  #
+  # When deploying, the proxy will by default hit /up once every second until we hit
+  # the deploy timeout, with a 5 second timeout for each request.
+  #
+  # Once the app is up, the proxy will stop hitting the healthcheck endpoint.
+  healthcheck:
+    interval: 3
+    path: /health
+    timeout: 3
+
+  # Buffering
+  #
+  # Whether to buffer request and response bodies in the proxy
+  #
+  # By default buffering is enabled with a max request body size of 1GB and no limit
+  # for response size.
+  #
+  # You can also set the memory limit for buffering, which defaults to 1MB, anything
+  # larger than that is written to disk.
+  buffering:
+    requests: true
+    responses: true
+    max_request_body: 40_000_000
+    max_response_body: 0
+    memory: 2_000_000
+
+  # Logging
+  #
+  # Configure request logging for the proxy
+  # You can specify request and response headers to log.
+  # By default, Cache-Control, Last-Modified and User-Agent request headers are logged
+  logging:
+    request_headers:
+      - Cache-Control
+      - X-Forwarded-Proto
+    response_headers:
+      - X-Request-ID
+      - X-Request-Start
+
+  # Forward headers
+  #
+  # Whether to forward the X-Forwarded-For and X-Forwarded-Proto headers.
+  #
+  # If you are behind a trusted proxy, you can set this to true to forward the headers.
+  #
+  # By default kamal-proxy will not forward the headers the ssl option is set to true, and
+  # will forward them if it is set to false.
+  forward_headers: true

lib/kamal/configuration/docs/registry.yml

@@ -27,11 +27,13 @@ registry:
 # and [set up roles and permissions](https://cloud.google.com/artifact-registry/docs/access-control#permissions).
 # Normally, assigning a roles/artifactregistry.writer role should be sufficient.
 #
-# Once the service account is ready, you need to generate and download a JSON key, base64 encode it and add to .env:
+# Once the service account is ready, you need to generate and download a JSON key and base64 encode it:
 #
 # ```shell
-# echo "KAMAL_REGISTRY_PASSWORD=$(base64 -i /path/to/key.json)" | tr -d "\\n"  >> .env
+# base64 -i /path/to/key.json | tr -d "\\n")
 # ```
+# You'll then need to set the KAMAL_REGISTRY_PASSWORD secret to that value.
+#
 # Use the env variable as password along with _json_key_base64 as username.
 # Here’s the final configuration:
 

lib/kamal/configuration/docs/role.yml

@@ -26,8 +26,12 @@ servers:
   #
   # When there are other options to set, the list of hosts goes under the `hosts` key
   #
-  # By default only the primary role uses Traefik, but you can set `traefik` to change
-  # it.
+  # By default only the primary role uses a proxy.
+  #
+  # For other roles, you can set it to `proxy: true` enable it and inherit the root proxy
+  # configuration or provide a map of options to override the root configuration.
+  #
+  # For the primary role, you can set `proxy: false` to disable the proxy.
   #
   # You can also set a custom cmd to run in the container, and overwrite other settings
   # from the root configuration.
@@ -35,18 +39,16 @@ servers:
     hosts:
       - 172.1.0.3
       - 172.1.0.4: experiment1
-    traefik: true
     cmd: "bin/jobs"
     options:
       memory: 2g
       cpus: 4
-    healthcheck:
-      ...
     logging:
       ...
+    proxy:
+      ...
     labels:
       my-label: workers
     env:
       ...
     asset_path: /public
-

lib/kamal/configuration/docs/traefik.yml

@@ -1,62 +0,0 @@
-# Traefik
-#
-# Traefik is a reverse proxy, used by Kamal for zero-downtime deployments.
-#
-# We start an instance on the hosts in it's own container.
-#
-# During a deployment:
-# 1. We start a new container which Traefik automatically detects due to the labels we have applied
-# 2. Traefik starts routing traffic to the new container
-# 3. We force the old container to fail it's healthcheck, causing Traefik to stop routing traffic to it
-# 4. We stop the old container
-
-# Traefik settings
-#
-# Traekik is configured in the root configuration under `traefik`.
-traefik:
-
-  # Image
-  #
-  # The Traefik image to use, defaults to `traefik:v2.10`
-  image: traefik:v2.9
-
-  # Host port
-  #
-  # The host port to publish the Traefik container on, defaults to `80`
-  host_port: "8080"
-
-  # Disabling publishing
-  #
-  # To avoid publishing the Traefik container, set this to `false`
-  publish: false
-
-  # Labels
-  #
-  # Additional labels to apply to the Traefik container
-  labels:
-    traefik.http.routers.catchall.entryPoints: http
-    traefik.http.routers.catchall.rule: PathPrefix(`/`)
-    traefik.http.routers.catchall.service: unavailable
-    traefik.http.routers.catchall.priority: "1"
-    traefik.http.services.unavailable.loadbalancer.server.port: "0"
-
-  # Arguments
-  #
-  # Additional arguments to pass to the Traefik container
-  args:
-    entryPoints.http.address: ":80"
-    entryPoints.http.forwardedHeaders.insecure: true
-    accesslog: true
-    accesslog.format: json
-
-  # Options
-  #
-  # Additional options to pass to `docker run`
-  options:
-    cpus: 2
-
-  # Environment variables
-  #
-  # See kamal docs env
-  env:
-    ...

lib/kamal/configuration/healthcheck.rb

@@ -1,63 +0,0 @@
-class Kamal::Configuration::Healthcheck
-  include Kamal::Configuration::Validation
-
-  attr_reader :healthcheck_config
-
-  def initialize(healthcheck_config:, context: "healthcheck")
-    @healthcheck_config = healthcheck_config || {}
-    validate! @healthcheck_config, context: context
-  end
-
-  def merge(other)
-    self.class.new healthcheck_config: healthcheck_config.deep_merge(other.healthcheck_config)
-  end
-
-  def cmd
-    healthcheck_config.fetch("cmd", http_health_check)
-  end
-
-  def port
-    healthcheck_config.fetch("port", 3000)
-  end
-
-  def path
-    healthcheck_config.fetch("path", "/up")
-  end
-
-  def max_attempts
-    healthcheck_config.fetch("max_attempts", 7)
-  end
-
-  def interval
-    healthcheck_config.fetch("interval", "1s")
-  end
-
-  def cord
-    healthcheck_config.fetch("cord", "/tmp/kamal-cord")
-  end
-
-  def log_lines
-    healthcheck_config.fetch("log_lines", 50)
-  end
-
-  def set_port_or_path?
-    healthcheck_config["port"].present? || healthcheck_config["path"].present?
-  end
-
-  def to_h
-    {
-      "cmd" => cmd,
-      "interval" => interval,
-      "max_attempts" => max_attempts,
-      "port" => port,
-      "path" => path,
-      "cord" => cord,
-      "log_lines" => log_lines
-    }
-  end
-
-  private
-    def http_health_check
-      "curl -f #{URI.join("http://localhost:#{port}", path)} || exit 1" if path.present? || port.present?
-    end
-end

lib/kamal/configuration/proxy.rb

@@ -0,0 +1,66 @@
+class Kamal::Configuration::Proxy
+  include Kamal::Configuration::Validation
+
+  DEFAULT_LOG_REQUEST_HEADERS = [ "Cache-Control", "Last-Modified", "User-Agent" ]
+  CONTAINER_NAME = "kamal-proxy"
+
+  delegate :argumentize, :optionize, to: Kamal::Utils
+
+  attr_reader :config, :proxy_config
+
+  def initialize(config:, proxy_config:, context: "proxy")
+    @config = config
+    @proxy_config = proxy_config
+    validate! @proxy_config, with: Kamal::Configuration::Validator::Proxy, context: context
+  end
+
+  def app_port
+    proxy_config.fetch("app_port", 80)
+  end
+
+  def ssl?
+    proxy_config.fetch("ssl", false)
+  end
+
+  def host
+    proxy_config["host"]
+  end
+
+  def deploy_options
+    {
+      host: proxy_config["host"],
+      tls: proxy_config["ssl"],
+      "deploy-timeout": seconds_duration(config.deploy_timeout),
+      "drain-timeout": seconds_duration(config.drain_timeout),
+      "health-check-interval": seconds_duration(proxy_config.dig("healthcheck", "interval")),
+      "health-check-timeout": seconds_duration(proxy_config.dig("healthcheck", "timeout")),
+      "health-check-path": proxy_config.dig("healthcheck", "path"),
+      "target-timeout": seconds_duration(proxy_config["response_timeout"]),
+      "buffer-requests": proxy_config.fetch("buffering", { "requests": true }).fetch("requests", true),
+      "buffer-responses": proxy_config.fetch("buffering", { "responses": true }).fetch("responses", true),
+      "buffer-memory": proxy_config.dig("buffering", "memory"),
+      "max-request-body": proxy_config.dig("buffering", "max_request_body"),
+      "max-response-body": proxy_config.dig("buffering", "max_response_body"),
+      "forward-headers": proxy_config.dig("forward_headers"),
+      "log-request-header": proxy_config.dig("logging", "request_headers") || DEFAULT_LOG_REQUEST_HEADERS,
+      "log-response-header": proxy_config.dig("logging", "response_headers")
+    }.compact
+  end
+
+  def deploy_command_args(target:)
+    optionize ({ target: "#{target}:#{app_port}" }).merge(deploy_options)
+  end
+
+  def remove_command_args(target:)
+    optionize({ target: "#{target}:#{app_port}" })
+  end
+
+  def merge(other)
+    self.class.new config: config, proxy_config: proxy_config.deep_merge(other.proxy_config)
+  end
+
+  private
+    def seconds_duration(value)
+      value ? "#{value}s" : nil
+    end
+end

lib/kamal/configuration/role.rb

@@ -1,10 +1,9 @@
 class Kamal::Configuration::Role
   include Kamal::Configuration::Validation
 
-  CORD_FILE = "cord"
   delegate :argumentize, :optionize, to: Kamal::Utils
 
-  attr_reader :name, :config, :specialized_env, :specialized_logging, :specialized_healthcheck
+  attr_reader :name, :config, :specialized_env, :specialized_logging, :specialized_proxy
 
   alias to_s name
 
@@ -25,9 +24,7 @@ class Kamal::Configuration::Role
       logging_config: specializations.fetch("logging", {}),
       context: "servers/#{name}/logging"
 
-    @specialized_healthcheck = Kamal::Configuration::Healthcheck.new \
-      healthcheck_config: specializations.fetch("healthcheck", {}),
-      context: "servers/#{name}/healthcheck"
+    initialize_specialized_proxy
   end
 
   def primary_host
@@ -55,7 +52,7 @@ class Kamal::Configuration::Role
   end
 
   def labels
-    default_labels.merge(traefik_labels).merge(custom_labels)
+    default_labels.merge(custom_labels)
   end
 
   def label_args
@@ -70,6 +67,24 @@ class Kamal::Configuration::Role
     @logging ||= config.logging.merge(specialized_logging)
   end
 
+  def proxy
+    @proxy ||= config.proxy.merge(specialized_proxy) if running_proxy?
+  end
+
+  def running_proxy?
+    @running_proxy
+  end
+
+  def ssl?
+    running_proxy? && proxy.ssl?
+  end
+
+  def stop_args
+    # When deploying with the proxy, kamal-proxy will drain request before returning so we don't need to wait.
+    timeout = running_proxy? ? nil : config.drain_timeout
+
+    [ *argumentize("-t", timeout) ]
+  end
 
   def env(host)
     @envs ||= {}
@@ -89,7 +104,7 @@ class Kamal::Configuration::Role
   end
 
   def secrets_path
-    File.join(config.env_directory, "roles", "#{container_prefix}.env")
+    File.join(config.env_directory, "roles", "#{name}.env")
   end
 
   def asset_volume_args
@@ -97,72 +112,8 @@ class Kamal::Configuration::Role
   end
 
 
-  def health_check_args(cord: true)
-    if running_traefik? || healthcheck.set_port_or_path?
-      if cord && uses_cord?
-        optionize({ "health-cmd" => health_check_cmd_with_cord, "health-interval" => healthcheck.interval })
-          .concat(cord_volume.docker_args)
-      else
-        optionize({ "health-cmd" => healthcheck.cmd, "health-interval" => healthcheck.interval })
-      end
-    else
-      []
-    end
-  end
-
-  def healthcheck
-    @healthcheck ||=
-      if running_traefik?
-        config.healthcheck.merge(specialized_healthcheck)
-      else
-        specialized_healthcheck
-      end
-  end
-
-  def health_check_cmd_with_cord
-    "(#{healthcheck.cmd}) && (stat #{cord_container_file} > /dev/null || exit 1)"
-  end
-
-
-  def running_traefik?
-    if specializations["traefik"].nil?
-      primary?
-    else
-      specializations["traefik"]
-    end
-  end
-
   def primary?
-    self == @config.primary_role
-  end
-
-
-  def uses_cord?
-    running_traefik? && cord_volume && healthcheck.cmd.present?
-  end
-
-  def cord_host_directory
-    File.join config.run_directory_as_docker_volume, "cords", [ container_prefix, config.run_id ].join("-")
-  end
-
-  def cord_volume
-    if (cord = healthcheck.cord)
-      @cord_volume ||= Kamal::Configuration::Volume.new \
-        host_path: File.join(config.run_directory, "cords", [ container_prefix, config.run_id ].join("-")),
-        container_path: cord
-    end
-  end
-
-  def cord_host_file
-    File.join cord_volume.host_path, CORD_FILE
-  end
-
-  def cord_container_directory
-    health_check_options.fetch("cord", nil)
-  end
-
-  def cord_container_file
-    File.join cord_volume.container_path, CORD_FILE
+    name == @config.primary_role_name
   end
 
 
@@ -180,25 +131,52 @@ class Kamal::Configuration::Role
   end
 
   def assets?
-    asset_path.present? && running_traefik?
+    asset_path.present? && running_proxy?
   end
 
-  def asset_volume(version = nil)
+  def asset_volume(version = config.version)
     if assets?
       Kamal::Configuration::Volume.new \
-        host_path: asset_volume_path(version), container_path: asset_path
+        host_path: asset_volume_directory(version), container_path: asset_path
     end
   end
 
-  def asset_extracted_path(version = nil)
-    File.join config.run_directory, "assets", "extracted", container_name(version)
+  def asset_extracted_directory(version = config.version)
+    File.join config.assets_directory, "extracted", [ name, version ].join("-")
   end
 
-  def asset_volume_path(version = nil)
-    File.join config.run_directory, "assets", "volumes", container_name(version)
+  def asset_volume_directory(version = config.version)
+    File.join config.assets_directory, "volumes", [ name, version ].join("-")
+  end
+
+  def ensure_one_host_for_ssl
+    if running_proxy? && proxy.ssl? && hosts.size > 1
+      raise Kamal::ConfigurationError, "SSL is only supported on a single server, found #{hosts.size} servers for role #{name}"
+    end
   end
 
   private
+    def initialize_specialized_proxy
+      proxy_specializations = specializations["proxy"]
+
+      if primary?
+        # only false means no proxy for non-primary roles
+        @running_proxy = proxy_specializations != false
+      else
+        # false and nil both mean no proxy for non-primary roles
+        @running_proxy = !!proxy_specializations
+      end
+
+      if running_proxy?
+        proxy_config = proxy_specializations == true || proxy_specializations.nil? ? {} : proxy_specializations
+
+        @specialized_proxy = Kamal::Configuration::Proxy.new \
+          config: config,
+          proxy_config: proxy_config,
+          context: "servers/#{name}/proxy"
+      end
+    end
+
     def tagged_hosts
       {}.tap do |tagged_hosts|
         extract_hosts_from_config.map do |host_config|
@@ -233,27 +211,6 @@ class Kamal::Configuration::Role
       end
     end
 
-    def traefik_labels
-      if running_traefik?
-        {
-          # Setting a service property ensures that the generated service name will be consistent between versions
-          "traefik.http.services.#{traefik_service}.loadbalancer.server.scheme" => "http",
-
-          "traefik.http.routers.#{traefik_service}.rule" => "PathPrefix(`/`)",
-          "traefik.http.routers.#{traefik_service}.priority" => "2",
-          "traefik.http.middlewares.#{traefik_service}-retry.retry.attempts" => "5",
-          "traefik.http.middlewares.#{traefik_service}-retry.retry.initialinterval" => "500ms",
-          "traefik.http.routers.#{traefik_service}.middlewares" => "#{traefik_service}-retry@docker"
-        }
-      else
-        {}
-      end
-    end
-
-    def traefik_service
-      container_prefix
-    end
-
     def custom_labels
       Hash.new.tap do |labels|
         labels.merge!(config.labels) if config.labels.present?

lib/kamal/configuration/traefik.rb

@@ -1,78 +0,0 @@
-class Kamal::Configuration::Traefik
-  delegate :argumentize, to: Kamal::Utils
-
-  DEFAULT_IMAGE = "traefik:v2.10"
-  CONTAINER_PORT = 80
-  DEFAULT_ARGS = {
-    "log.level" => "DEBUG"
-  }
-  DEFAULT_LABELS = {
-    # These ensure we serve a 502 rather than a 404 if no containers are available
-    "traefik.http.routers.catchall.entryPoints" => "http",
-    "traefik.http.routers.catchall.rule" => "PathPrefix(`/`)",
-    "traefik.http.routers.catchall.service" => "unavailable",
-    "traefik.http.routers.catchall.priority" => 1,
-    "traefik.http.services.unavailable.loadbalancer.server.port" => "0"
-  }
-
-  include Kamal::Configuration::Validation
-
-  attr_reader :config, :traefik_config
-
-  def initialize(config:)
-    @config = config
-    @traefik_config = config.raw_config.traefik || {}
-    validate! traefik_config
-  end
-
-  def publish?
-    traefik_config["publish"] != false
-  end
-
-  def labels
-    DEFAULT_LABELS.merge(traefik_config["labels"] || {})
-  end
-
-  def env
-    Kamal::Configuration::Env.new \
-      config: traefik_config.fetch("env", {}),
-      secrets: config.secrets,
-      context: "traefik/env"
-  end
-
-  def host_port
-    traefik_config.fetch("host_port", CONTAINER_PORT)
-  end
-
-  def options
-    traefik_config.fetch("options", {})
-  end
-
-  def port
-    "#{host_port}:#{CONTAINER_PORT}"
-  end
-
-  def args
-    DEFAULT_ARGS.merge(traefik_config.fetch("args", {}))
-  end
-
-  def image
-    traefik_config.fetch("image", DEFAULT_IMAGE)
-  end
-
-  def env_args
-    [ *env.clear_args, *argumentize("--env-file", secrets_path) ]
-  end
-
-  def env_directory
-    File.join(config.env_directory, "traefik")
-  end
-
-  def secrets_io
-    env.secrets_io
-  end
-
-  def secrets_path
-    File.join(config.env_directory, "traefik", "traefik.env")
-  end
-end

lib/kamal/configuration/validator.rb

@@ -24,7 +24,9 @@ class Kamal::Configuration::Validator
             example_value = example[key]
 
             if example_value == "..."
-              validate_type! value, *(Array if key == :servers), Hash
+              unless key.to_s == "proxy" && boolean?(value.class)
+                validate_type! value, *(Array if key == :servers), Hash
+              end
             elsif key == "hosts"
               validate_servers! value
             elsif example_value.is_a?(Array)

lib/kamal/configuration/validator/proxy.rb

@@ -0,0 +1,11 @@
+class Kamal::Configuration::Validator::Proxy < Kamal::Configuration::Validator
+  def validate!
+    unless config.nil?
+      super
+
+      if config["host"].blank? && config["ssl"]
+        error "Must set a host to enable automatic SSL"
+      end
+    end
+  end
+end

lib/kamal/utils.rb

@@ -101,4 +101,8 @@ module Kamal::Utils
       arch
     end
   end
+
+  def older_version?(version, other_version)
+    Gem::Version.new(version.delete_prefix("v")) < Gem::Version.new(other_version.delete_prefix("v"))
+  end
 end

lib/kamal/version.rb

@@ -1,3 +1,3 @@
 module Kamal
-  VERSION = "2.0.0.alpha"
+  VERSION = "2.0.0.rc2"
 end