Skip to content

fix(operator): rollout-restart failures are swallowed — CR still reports Ready=True #214

Description

@BryanFRD

Found during the 2026-08 FerrVault operator audit.

Problem

When triggerRollouts returns an error it is logged and discarded (internal/controller/ferrvaultsecret_controller.go:147-151); the reconcile still proceeds to set Ready=True and returns success. A user whose dependent workloads never restart (e.g. the 403 in the rollout-verb issue, or a bad WorkloadRef) sees a green FerrVaultSecret with no signal that the rollout failed, and no metric increments.

Fix

Surface rollout failures on a distinct status condition (e.g. RolloutSucceeded=False) and/or increment a SyncErrors{reason="RolloutFailed"} metric, so kubectl describe and alerting reflect it.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    P2Medium prioritybugSomething isn't working

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions