Operational runbooks for the addon catalog#86
Merged
Merged
Conversation
docs/runbooks/ carried a single troubleshooting reference for a repo that
deploys ~40 addons across four environments. This adds four operational
runbooks, each grounded in what the repo actually ships (applicationsets,
the druid catalog chart, the CI render gates) and shaped
symptoms → diagnosis → remediation → verification.
─── addon-sync-degraded ───
Diagnosing an ArgoCD Application that is stuck, degraded, or missing:
reading status.conditions/operationState, why a missing cluster-secret
environment label means no app at all (missingkey=error), sync-wave
implications for CRD-before-consumer ordering, the ignoreDifferences
contract for StatefulSet volumeClaimTemplates, and targeted
hard-refresh / per-resource sync / terminate-op levers — with the
explicit caveat that the ApplicationSet controller owns the generated
Application spec, so manual overrides are transient by design.
─── rollback ───
Chart-pin revert flow: pins live in the ApplicationSet list elements,
values deltas in addons/<category>/<addon>/, and because every app runs
automated {prune, selfHeal}, the only durable rollback is a git revert
through the same render gates as the original change. Documents the
temporary incident levers (sync-policy none, app rollback) and their
self-heal expiry, plus version-specific gotchas: CRD-managing charts
don't downgrade CRDs, StatefulSet volumeClaimTemplates are immutable.
─── druid-operations ───
The chart this repo owns outright. Keystore secret rotation end to end:
AWS Secrets Manager → ExternalSecret (1h refresh, force-sync
annotation) → cert-manager passwordSecretRef re-encryption → pod
restart, with the ordering constraint that pods must not restart
between the password sync and the keystore re-issue. Probe semantics
table (per-component ports, the broker/historical readiness endpoints
vs plain /status/health, the ~11-minute startup window, why HTTPS
probes work against the self-signed internal CA). Scaling guidance per
component class: historicals gated on segment reload from deep storage,
stateless broker/router, leader-elected coordinator/overlord, and
ingestion capacity living in the overlord's Job launcher rather than
replicas — all through git because selfHeal reverts kubectl scale.
─── render-gate-failures ───
What each CI gate failure means (zero-placeholder, render-assert,
kubeconform strict with no ignore-missing-schemas, trivy config
MEDIUM+) and the local task equivalents to reproduce. Codifies the
exception discipline: a .trivyignore.yaml entry is legitimate only when
the finding is the component's contract, must be path-scoped to one
addon's rendered output, and must carry a statement that survives a
cold review — the existing device-plugin and druid entries are the
calibration. Same per-kind, commented discipline for kubeconform skips.
README's documentation index groups the runbooks under a Runbooks list.
Co-authored-by: stxkxsbot <275011021+stxkxsbot@users.noreply.github.com>
CI Results
All checks passed. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
See commit message. Four runbooks (addon-sync-degraded, rollback, druid-operations, render-gate-failures), each grounded in the applicationsets/catalog/CI; symptoms→diagnosis→remediation→verification shape.