Skip to content

fix: prevent renewal retry storms and handle orphaned Valid orders #6

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

Merged
wgordon17 merged 2 commits into from
Sep 25, 2026
Merged

fix: prevent renewal retry storms and handle orphaned Valid orders #6

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
3565793
fix(route): fix exponential backoff math and handle orphaned Valid or…
wgordon17 Sep 25, 2026
f4bc177
fix(route): use RetryOnConflict when attaching cert to route
wgordon17 Sep 25, 2026
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
47 changes: 39 additions & 8 deletions pkg/controller/route/route.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -40,6 +40,7 @@ import (
typedcorev1 "k8s.io/client-go/kubernetes/typed/core/v1"
"k8s.io/client-go/tools/cache"
"k8s.io/client-go/tools/record"
"k8s.io/client-go/util/retry"
"k8s.io/client-go/util/workqueue"
"k8s.io/klog/v2"

Expand Down Expand Up @@ -1134,10 +1135,28 @@ func (rc *RouteController) sync(ctx context.Context, key string) error {
route.Spec.TLS.Key = string(certPemData.Key)
route.Spec.TLS.Certificate = string(certPemData.Crt)

// TODO: consider RetryOnConflict with rechecking the managed annotation
_, err = rc.routeClient.RouteV1().Routes(routeReadOnly.Namespace).Update(ctx, route, metav1.UpdateOptions{})
// Use RetryOnConflict to prevent dropping the in-memory private key and
// burning Let's Encrypt rate limits on transient Kubernetes update conflicts.
err = retry.RetryOnConflict(retry.DefaultRetry, func() error {
freshRoute, getErr := rc.routeClient.RouteV1().Routes(routeReadOnly.Namespace).Get(ctx, routeReadOnly.Name, metav1.GetOptions{})
if getErr != nil {
return getErr
}

if freshRoute.Spec.TLS == nil {
freshRoute.Spec.TLS = &routev1.TLSConfig{
InsecureEdgeTerminationPolicy: routev1.InsecureEdgeTerminationPolicyRedirect,
Termination: routev1.TLSTerminationEdge,
}
}
freshRoute.Spec.TLS.Key = string(certPemData.Key)
freshRoute.Spec.TLS.Certificate = string(certPemData.Crt)

_, updateErr := rc.routeClient.RouteV1().Routes(freshRoute.Namespace).Update(ctx, freshRoute, metav1.UpdateOptions{})
return updateErr
})
if err != nil {
return fmt.Errorf("can't update route %s/%s with new certificates: %v", routeReadOnly.Namespace, route.Name, err)
return fmt.Errorf("can't update route %s/%s with new certificates after retries: %v", routeReadOnly.Namespace, routeReadOnly.Name, err)
}

err = rc.CleanupExposerObjects(ctx, routeReadOnly)
Expand All @@ -1149,11 +1168,23 @@ func (rc *RouteController) sync(ctx context.Context, key string) error {
return nil

case acme.StatusValid:
// TODO: fix the golang acme lib
// Unfortunately the golang acme lib actively waits in 'CreateOrderCert'
// so we can't take the appropriate asynchronous action here.
// The logic is included in handling acme.StatusReady
return nil
// We have a Valid order on the ACME server, but we lost the in-memory private key
// needed to attach the certificate (likely due to a previous Route Update conflict).
// We must abandon this orphaned order, trigger a backoff, and start fresh.
klog.Warningf("Route %q: Order %q is Valid but we lack the private key. Restarting order.", key, order.URI)

if status.ProvisioningStatus.OrderStatus != previousOrderStatus {
status.ProvisioningStatus.Failures += 1
}

err = rc.CleanupExposerObjects(ctx, routeReadOnly)
if err != nil {
klog.Errorf("Can't cleanup exposer objects: %v", err)
}

status.ProvisioningStatus.OrderURI = ""
status.ProvisioningStatus.OrderStatus = ""
return rc.updateStatus(ctx, routeReadOnly, status)

case acme.StatusInvalid:
rc.recorder.Eventf(routeReadOnly, corev1.EventTypeWarning, "AcmeFailedOrder", "Order %q for domain %q failed: %v", order.URI, routeReadOnly.Spec.Host, order.Error)
Expand Down
Loading