Merge pull request #1392 from fluxcd/reduce-mttr

Reduce the mean time to recovery (MTTR) in case of failed deployments

Author: matheuscscp

.github/workflows/e2e.yaml

@@ -712,8 +712,6 @@ jobs:
 
           kubectl -n helm-system delete -f config/testdata/$test_name
       - name: Run install test fail test
-        # TODO: re-enable after upgrading to Helm 4.1.1 (https://github.com/helm/helm/pull/31730)
-        if: false
         run: |
           test_name=install-test-fail
           kubectl -n helm-system apply -f config/testdata/$test_name
@@ -739,8 +737,6 @@ jobs:
 
           kubectl -n helm-system delete -f config/testdata/$test_name
       - name: Run install test fail ignore test
-        # TODO: re-enable after upgrading to Helm 4.1.1 (https://github.com/helm/helm/pull/31730)
-        if: false
         run: |
           test_name=install-test-fail-ignore
           kubectl -n helm-system apply -f config/testdata/$test_name
@@ -863,8 +859,6 @@ jobs:
 
           kubectl delete -n helm-system -f config/testdata/$test_name/install.yaml
       - name: Run upgrade test fail test
-        # TODO: re-enable after upgrading to Helm 4.1.1 (https://github.com/helm/helm/pull/31730)
-        if: false
         run: |
           test_name=upgrade-test-fail
           kubectl -n helm-system apply -f config/testdata/$test_name/install.yaml

api/v2/helmrelease_types.go

@@ -194,7 +194,7 @@ type HelmReleaseSpec struct {
 	// health of custom resources using Common Expression Language (CEL).
 	// The expressions are evaluated only when the specific Helm action
 	// taking place has wait enabled, i.e. DisableWait is false, and the
-	// 'watcher' WaitStrategy is used.
+	// 'poller' WaitStrategy is used.
 	// +optional
 	HealthCheckExprs []kustomize.CustomHealthCheck `json:"healthCheckExprs,omitempty"`
 }
@@ -441,8 +441,8 @@ type HelmChartTemplateVerification struct {
 type WaitStrategyName string
 
 const (
-	// WaitStrategyWatcher is the strategy for watching resource statuses via kstatus.
-	WaitStrategyWatcher WaitStrategyName = "watcher"
+	// WaitStrategyPoller is the strategy for polling resource statuses via kstatus.
+	WaitStrategyPoller WaitStrategyName = "poller"
 
 	// WaitStrategyLegacy is the legacy strategy for waiting for resources to be ready
 	// used in Helm v3.
@@ -453,12 +453,12 @@ const (
 // resources to become ready.
 type WaitStrategy struct {
 	// Name is Helm's wait strategy for waiting for applied resources to
-	// become ready. One of 'watcher' or 'legacy'. The 'watcher' strategy uses
-	// kstatus to watch resource statuses, while the 'legacy' strategy uses
+	// become ready. One of 'poller' or 'legacy'. The 'poller' strategy uses
+	// kstatus to poll resource statuses, while the 'legacy' strategy uses
 	// Helm v3's waiting logic.
-	// Defaults to 'watcher', or to 'legacy' when UseHelm3Defaults feature
+	// Defaults to 'poller', or to 'legacy' when UseHelm3Defaults feature
 	// gate is enabled.
-	// +kubebuilder:validation:Enum=watcher;legacy
+	// +kubebuilder:validation:Enum=poller;legacy
 	// +required
 	Name WaitStrategyName `json:"name"`
 }

config/crd/bases/helm.toolkit.fluxcd.io_helmreleases.yaml

@@ -355,7 +355,7 @@ spec:
                   health of custom resources using Common Expression Language (CEL).
                   The expressions are evaluated only when the specific Helm action
                   taking place has wait enabled, i.e. DisableWait is false, and the
-                  'watcher' WaitStrategy is used.
+                  'poller' WaitStrategy is used.
                 items:
                   description: CustomHealthCheck defines the health check for custom
                     resources.
@@ -1102,13 +1102,13 @@ spec:
                   name:
                     description: |-
                       Name is Helm's wait strategy for waiting for applied resources to
-                      become ready. One of 'watcher' or 'legacy'. The 'watcher' strategy uses
-                      kstatus to watch resource statuses, while the 'legacy' strategy uses
+                      become ready. One of 'poller' or 'legacy'. The 'poller' strategy uses
+                      kstatus to poll resource statuses, while the 'legacy' strategy uses
                       Helm v3's waiting logic.
-                      Defaults to 'watcher', or to 'legacy' when UseHelm3Defaults feature
+                      Defaults to 'poller', or to 'legacy' when UseHelm3Defaults feature
                       gate is enabled.
                     enum:
-                    - watcher
+                    - poller
                     - legacy
                     type: string
                 required:

config/testdata/server-side-apply/rollback-upgrade.yaml

@@ -29,4 +29,4 @@ spec:
     replicaCount: 2
     faults:
       unready: true
-  timeout: 3s
+  timeout: 10s

config/testdata/upgrade-fail-remediate/upgrade.yaml

@@ -25,4 +25,4 @@ spec:
     replicaCount: 2
     faults:
       unready: true
-  timeout: 3s
+  timeout: 10s

config/testdata/upgrade-fail-retry/upgrade.yaml

@@ -25,4 +25,4 @@ spec:
     replicaCount: 2
     faults:
       unready: true
-  timeout: 3s
+  timeout: 10s

docs/api/v2/helm.md

@@ -435,7 +435,7 @@ resources to become ready.</p>
 health of custom resources using Common Expression Language (CEL).
 The expressions are evaluated only when the specific Helm action
 taking place has wait enabled, i.e. DisableWait is false, and the
-&lsquo;watcher&rsquo; WaitStrategy is used.</p>
+&lsquo;poller&rsquo; WaitStrategy is used.</p>
 </td>
 </tr>
 </table>
@@ -1603,7 +1603,7 @@ resources to become ready.</p>
 health of custom resources using Common Expression Language (CEL).
 The expressions are evaluated only when the specific Helm action
 taking place has wait enabled, i.e. DisableWait is false, and the
-&lsquo;watcher&rsquo; WaitStrategy is used.</p>
+&lsquo;poller&rsquo; WaitStrategy is used.</p>
 </td>
 </tr>
 </tbody>
@@ -3471,10 +3471,10 @@ WaitStrategyName
 </td>
 <td>
 <p>Name is Helm&rsquo;s wait strategy for waiting for applied resources to
-become ready. One of &lsquo;watcher&rsquo; or &lsquo;legacy&rsquo;. The &lsquo;watcher&rsquo; strategy uses
-kstatus to watch resource statuses, while the &lsquo;legacy&rsquo; strategy uses
+become ready. One of &lsquo;poller&rsquo; or &lsquo;legacy&rsquo;. The &lsquo;poller&rsquo; strategy uses
+kstatus to poll resource statuses, while the &lsquo;legacy&rsquo; strategy uses
 Helm v3&rsquo;s waiting logic.
-Defaults to &lsquo;watcher&rsquo;, or to &lsquo;legacy&rsquo; when UseHelm3Defaults feature
+Defaults to &lsquo;poller&rsquo;, or to &lsquo;legacy&rsquo; when UseHelm3Defaults feature
 gate is enabled.</p>
 </td>
 </tr>

docs/spec/v2/helmreleases.md

@@ -952,14 +952,14 @@ for resources to become ready after Helm actions.
 The field offers the following subfields:
 
 - `.name` (Required): The strategy for waiting for resources to be ready.
-  One of `watcher` or `legacy`. The `watcher` strategy uses kstatus to watch resource
+  One of `poller` or `legacy`. The `poller` strategy uses kstatus to poll resource
   statuses, while the `legacy` strategy uses Helm v3's waiting logic. Defaults to
-  `watcher`, or to `legacy` when the `UseHelm3Defaults` feature gate is enabled.
+  `poller`, or to `legacy` when the `UseHelm3Defaults` feature gate is enabled.
 
 ```yaml
 spec:
   waitStrategy:
-    name: watcher
+    name: poller
 ```
 
 ### Health check expressions
@@ -968,8 +968,8 @@ spec:
 checks on custom resources using [Common Expression Language (CEL)](https://cel.dev/).
 
 The expressions are evaluated only when the Helm action taking place has wait
-enabled (i.e. `.spec.<action>.disableWait` is `false`) and the `watcher`
-wait strategy is used (i.e. `.spec.waitStrategy.name` is `watcher`).
+enabled (i.e. `.spec.<action>.disableWait` is `false`) and the `poller`
+wait strategy is used (i.e. `.spec.waitStrategy.name` is `poller`).
 
 The `.spec.healthCheckExprs` field accepts a list of objects with the following fields:
 

go.mod

@@ -25,8 +25,8 @@ require (
 	github.com/fluxcd/pkg/auth v0.36.0
 	github.com/fluxcd/pkg/cache v0.13.0
 	github.com/fluxcd/pkg/chartutil v1.21.0
-	github.com/fluxcd/pkg/runtime v0.96.0
-	github.com/fluxcd/pkg/ssa v0.64.0
+	github.com/fluxcd/pkg/runtime v0.97.0
+	github.com/fluxcd/pkg/ssa v0.65.0
 	github.com/fluxcd/pkg/testserver v0.13.0
 	github.com/fluxcd/source-controller/api v1.7.2
 	github.com/go-logr/logr v1.4.3

go.sum

@@ -160,10 +160,10 @@ github.com/fluxcd/pkg/cache v0.13.0 h1:MqtlgOwIVcGKKgV422e39O+KFSVMWuExKeRaMDBjJ
 github.com/fluxcd/pkg/cache v0.13.0/go.mod h1:0xRZ1hitrIFQ6pl68ke2wZLbIqA2VLzY78HpDo9DVxs=
 github.com/fluxcd/pkg/chartutil v1.21.0 h1:NJYhlekwBwuqMpRgsOlcsJrw2Xq0cBJW0Nmvz2oMluA=
 github.com/fluxcd/pkg/chartutil v1.21.0/go.mod h1:Gv50bF3SS4OvvKCyyIMRkGeNzZk6Fsh4+lAdrjx97T4=
-github.com/fluxcd/pkg/runtime v0.96.0 h1:sF4ic8131BwbOE+T2pkiXlkr2gCaxAho500zlZJJLck=
-github.com/fluxcd/pkg/runtime v0.96.0/go.mod h1:FyjNMFNAERkCsF/muTWJYU9MZOsq/m4Sc4aQk/EgQ9E=
-github.com/fluxcd/pkg/ssa v0.64.0 h1:B/8VYMIYMeRmolup2HOoWNqXh4UeXi6w2LvXXvl6MZM=
-github.com/fluxcd/pkg/ssa v0.64.0/go.mod h1:RjvVjJIoRo1ecsv91yMuiqzO6cpNag80M6MOB/vrJdc=
+github.com/fluxcd/pkg/runtime v0.97.0 h1:NZ9ZL2RE6J4yNDHI7ZNmL/OnCsDWurzVIvws73FYMz8=
+github.com/fluxcd/pkg/runtime v0.97.0/go.mod h1:FyjNMFNAERkCsF/muTWJYU9MZOsq/m4Sc4aQk/EgQ9E=
+github.com/fluxcd/pkg/ssa v0.65.0 h1:pvPYmBmfqS0sRNoi4/IUoocBbglYuD/OJLgOuxMJ9+o=
+github.com/fluxcd/pkg/ssa v0.65.0/go.mod h1:RjvVjJIoRo1ecsv91yMuiqzO6cpNag80M6MOB/vrJdc=
 github.com/fluxcd/pkg/testserver v0.13.0 h1:xEpBcEYtD7bwvZ+i0ZmChxKkDo/wfQEV3xmnzVybSSg=
 github.com/fluxcd/pkg/testserver v0.13.0/go.mod h1:akRYv3FLQUsme15na9ihECRG6hBuqni4XEY9W8kzs8E=
 github.com/fluxcd/source-controller/api v1.7.2 h1:/lg/xoyRjxwdhHKqjTxQS2o1cp+DMKJ8W4rpm+ZLemQ=