Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[processor/k8sattributes] Wait for ReplicaSet informer before starting pod informer #37138

Draft
wants to merge 2 commits into
base: main
Choose a base branch
from
+29 −6
Draft

[processor/k8sattributes] Wait for ReplicaSet informer before starting pod informer #37138

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
de3aad2
wait for replica set informer before starting pod informer
bacherfl Jan 10, 2025
b4d7636
re-enable failing test
bacherfl Jan 10, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 0 additions & 2 deletions processor/k8sattributesprocessor/e2e_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1099,8 +1099,6 @@ func TestE2E_NamespacedRBACNoPodIP(t *testing.T) {
// make docker-otelcontribcol
// KUBECONFIG=/tmp/kube-config-otelcol-e2e-testing kind load docker-image otelcontribcol:latest
func TestE2E_ClusterRBACCollectorStartAfterTelemetryGen(t *testing.T) {
// TODO: Re-enable this test when the issue being tested here is fully solved: https://github.com/open-telemetry/opentelemetry-collector-contrib/issues/37056
t.Skip("Skipping test as https://github.com/open-telemetry/opentelemetry-collector-contrib/issues/37056 is not fully solved yet")
testDir := filepath.Join("testdata", "e2e", "clusterrbac")

k8sClient, err := k8stest.NewK8sClient(testKubeConfig)
Expand Down
33 changes: 29 additions & 4 deletions processor/k8sattributesprocessor/internal/kube/client.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -210,9 +210,11 @@ func New(
func (c *WatchClient) Start() error {
synced := make([]cache.InformerSynced, 0)

waitForReplicaSets := false
// start the replicaSet informer first, as the replica sets need to be
// present at the time the pods are handled, to correctly establish the connection between pods and deployments
if c.Rules.DeploymentName || c.Rules.DeploymentUID {
waitForReplicaSets = true
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Waiting for the ReplicaSet informer to sync before allowing Pod informer to sync make sense.

Should we consider waiting for other resources too before unblocking the Pod's sync?

I wonder if it is just cleaner to start all the informers in parallel except of the Pod informer, wait for them to sync and then start the Pod informer eventually ensuring that all the dependencies are in sync.

reg, err := c.replicasetInformer.AddEventHandler(cache.ResourceEventHandlerFuncs{
AddFunc: c.handleReplicaSetAdd,
UpdateFunc: c.handleReplicaSetUpdate,
Expand All @@ -222,7 +224,7 @@ func (c *WatchClient) Start() error {
return err
}
synced = append(synced, reg.HasSynced)
go c.replicasetInformer.Run(c.stopCh)
go c.runInformer(c.replicasetInformer)
}

reg, err := c.informer.AddEventHandler(cache.ResourceEventHandlerFuncs{
Expand All @@ -234,7 +236,12 @@ func (c *WatchClient) Start() error {
return err
}
synced = append(synced, reg.HasSynced)
go c.informer.Run(c.stopCh)

if waitForReplicaSets {
go c.runInformer(c.informer, c.replicasetInformer)
} else {
go c.runInformer(c.informer)
}

reg, err = c.namespaceInformer.AddEventHandler(cache.ResourceEventHandlerFuncs{
AddFunc: c.handleNamespaceAdd,
Expand All @@ -245,7 +252,7 @@ func (c *WatchClient) Start() error {
return err
}
synced = append(synced, reg.HasSynced)
go c.namespaceInformer.Run(c.stopCh)
go c.runInformer(c.namespaceInformer)

if c.nodeInformer != nil {
reg, err = c.nodeInformer.AddEventHandler(cache.ResourceEventHandlerFuncs{
Expand All @@ -257,7 +264,7 @@ func (c *WatchClient) Start() error {
return err
}
synced = append(synced, reg.HasSynced)
go c.nodeInformer.Run(c.stopCh)
go c.runInformer(c.nodeInformer)
}

if c.waitForMetadata {
Expand Down Expand Up @@ -1123,6 +1130,24 @@ func (c *WatchClient) getReplicaSet(uid string) (*ReplicaSet, bool) {
return nil, false
}

// runInformer starts the given informer. This method optionally takes a list of other informers that should complete
// before the informer is started. This is necessary e.g. for the pod informer which requires the replica set informer
// to be finished to correctly establish the connection to the replicaset/deployment it belongs to.
func (c *WatchClient) runInformer(informer cache.SharedInformer, dependencies ...cache.SharedInformer) {
if len(dependencies) > 0 {
timeoutCh := make(chan struct{})
// TODO hard coding the timeout for now, check if we should make this configurable
t := time.AfterFunc(5*time.Second, func() {
close(timeoutCh)
})
defer t.Stop()
for _, dependency := range dependencies {
cache.WaitForCacheSync(timeoutCh, dependency.HasSynced)
}
}
informer.Run(c.stopCh)
}

// ignoreDeletedFinalStateUnknown returns the object wrapped in
// DeletedFinalStateUnknown. Useful in OnDelete resource event handlers that do
// not need the additional context.
Expand Down
Loading