fix(release): delete a retryably-Failed evidence exporter so the deferred export can converge #2047
Loading…
Reference in a new issue
No description provided.
Delete branch "fix/export-retry-adoption"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
The signer adoption wedge, one obligation later: reconcileEvidenceExport released the latched name on Failed, but the Failed workflow kept its deterministic name and the paced resubmission ADOPTED it. All six deferred evidence exports are frozen this way live right now — without this, tonight's quota GC clears and nothing converges.
Fix mirrors PR #2046: delete the aged Failed exporter (finishedAt older than evidenceRetryDelay, via argoengine.Times) before releasing the name; a fresh failure stays latched until the delay passes. The envtest used to perform the manual unwedge itself (k8sClient.Delete before recovery) — it now proves the operator needs no hand. DOMAIN-MODEL durable-evidence caveat updated to the true split (release export self-heals; deployment audit manual-by-design).
130/130 guards, full controller suite green. Once deployed, the operator un-freezes the six wedged exports on its own.