controllers: use single error handling for all CRs

[AndrewChubatiuk]

Summary by cubic

Unifies error handling across all controllers and standardizes how status metadata is read from CRs. Adds graceful shutdown handling and better error propagation while keeping NotFound silent and requeuing on cancellations and conflicts.

• Refactors

  • Removed handleReconcileErrWithoutStatus; all controllers now use a single handleReconcileErr.
  • Moved GetStatusMetadata from Status structs to parent CRs; updated helpers (waitForStatus, status updates) to read metadata from the CR.
  • Added GetStatus/DefaultStatusFields for VMAlertmanagerConfig, VMRule, VMUser, and scrape CRs (VMNodeScrape, VMPodScrape, VMProbe, VMScrapeConfig, VMServiceScrape, VMStaticScrape); standardized controllers to use value instances and pass pointers.

• Bug Fixes

  • Requeue on context.Canceled unless caused by graceful shutdown (operator.ErrShutdown); keep NotFound silent and back off on conflicts.
  • Emit events only on real errors; consistent error metrics across all CRs.
  • Stop reconciliation on invalid specs via spec.ParsingError; VMRule and VMUser capture JSON parse errors during unmarshal.
  • Propagate errors from related-resource updates while continuing other items (VMAgent/VMSingle scrapes, alert/rule config updates).

^{Written for commit 551ba04. Summary will update on new commits.}

[cubic-dev-ai]

4 issues found across 49 files

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="api/operator/v1beta1/vmrule_types.go">

<violation number="1" location="api/operator/v1beta1/vmrule_types.go:249">
P1: Returning nil here hides VMRule JSON decode failures from admission; malformed rules can now be accepted and only fail later in reconciliation.</violation>
</file>

<file name="internal/controller/operator/vmrule_controller.go">

<violation number="1" location="internal/controller/operator/vmrule_controller.go:122">
P1: Don't only log configmap update failures here. A later successful iteration clears the shared named `err`, so this reconcile can return success after partially failing to update `VMAlert` configmaps.</violation>
</file>

<file name="internal/controller/operator/vmuser_controller.go">

<violation number="1" location="internal/controller/operator/vmuser_controller.go:131">
P1: Propagate this `CreateOrUpdateConfig` failure instead of only logging it; otherwise the reconcile returns success and controller-runtime will not retry the VMUser update.</violation>
</file>

<file name="internal/controller/operator/vmalertmanagerconfig_controller.go">

<violation number="1" location="internal/controller/operator/vmalertmanagerconfig_controller.go:75">
P2: Treat `ParsingError` as non-retryable here; returning it through `handleReconcileErr` will put malformed configs into a permanent requeue/event loop.</violation>
</file>

_{Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review.}

[cubic-dev-ai]

1 issue found across 2 files (changes from recent commits).

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="internal/controller/operator/controllers.go">

<violation number="1" location="internal/controller/operator/controllers.go:133">
P2: Check `context.Cause(ctx)` here instead of `err`; `WithCancelCause` stores `ErrShutdown` on the context, so graceful shutdowns will usually still be requeued.</violation>
</file>

_{Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review.}

[vrutkovs]

Why not resultErr? Not sure if we want to emit a new event on every GET failure and parsing error

[AndrewChubatiuk]

removed resultErr, initially it was added for a different reason