The liveness probe used the default timeoutSeconds of 1. The auto-triage tick
runs every minute and spends most of it waiting on Jenkins, OpenSearch, Gitea
and Hermes, so against a 500m CPU limit /health occasionally answers in over a
second. Three of those and the container is killed, dropping triage ticks for
the length of a restart. Observed 11 times in 139 minutes, with the pod
sitting 1/2 Ready and restarting repeatedly.
Give both probes a 5s timeout and let liveness tolerate five failures, so a
busy tick is no longer mistaken for a hung process.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The Codex weekly limit is close, so make anthropic/claude-opus-5 primary and
demote openai-codex to first fallback with the local gpt-oss:20b behind it.
The credential is a Claude subscription OAuth token, not an API key. The
anthropic provider resolves ANTHROPIC_API_KEY, then ANTHROPIC_TOKEN, then
CLAUDE_CODE_OAUTH_TOKEN, so the OAuth token must arrive under the last name
to be treated correctly. Marked optional so Hermes still starts and falls
back if the Secret is absent.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
ariadne PR #3 from incident ariadne/409, chosen path ariadne/utils/errors.py -
a file that appears nowhere in the build console and was reached only by
following the failing test's imports.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Every human_required escalation already files an issue in the failing
service's own repository, and mailing on each one made the inbox the loudest
and least informative output of the system. Replace the blanket alert with
two narrow ones: a repair that ran and failed, which is the only case where
the automation acted and left things no better, and an escalation still
untouched after six hours.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Red-build to green-rebuild is 1m04s on Ariadne 0.1.0-402, down from 4m00s
armed-to-resolved when the repair spawned its own Kubernetes Job. Reframe the
timings around the red build, since the arming leg depends on the agent pool.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Two conditions silently break a rehearsal. A saturated Kubernetes agent pool
leaves the demo build queued reporting that all nodes are offline, and an open
hermes-repair PR makes the duplicate guard refuse a new proposal. Report both
so the operator sees them before starting rather than mid-demo.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The classification->action registry already mapped transient_infra_failure to
retry_transient_infra, but the action allowlist held only repair_demo_fixture,
so that route always died at the action_not_allowlisted gate. Add the action
and state the registry explicitly rather than relying on the code default.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The webhook is failurePolicy: Ignore, so with one replica any pod created
during an injector restart is admitted without its Vault agent sidecar and
then crash-loops forever on a missing /vault/secrets file, with nothing to
indicate injection was skipped. Hit twice while rolling ariadne.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The fixture repair now runs in-process inside Ariadne, so the
hermes-demo-repair service account and its Role/RoleBinding have no
remaining user.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
ROOT still used parents[1], which resolved to scripts/ once the renderer
moved into scripts/render/. Every --build run wrote a phantom
scripts/services/monitoring tree and silently left the real dashboards
untouched. Points at the repo root again and removes the stray tree.
Also adds Hermes triage panels to the Atlas Testing dashboard: open
escalations awaiting a human, automated actions succeeded, Hermes
diagnosis latency, actions by result, and incident state by job.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- Ariadne: per-repo code config for metis, lesavka, soteria,
bstein-dev-home and ariadne, each with its own base branch, source path
prefixes and file suffixes so a proposal can only touch that repo's
source tree
- Alertmanager: the only receiver was an empty "default", so every alert
fired into a void. HermesTriageHumanRequired now routes to an email
receiver via Mailu's in-cluster local-domain relay, with resolved
notices; scoped to service=hermes-triage so nothing else mails yet.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Adds ariadne, lesavka, soteria, metis, ananke, pegasus, atlasbot,
bstein-dev-home and data-prepper to the autotriage allowlist. These get
detection, evidence, Hermes diagnosis and human-required escalation only:
no allowlisted repair action applies to them, so nothing is mutated.
Verified on lesavka #582 with the new console failure-region evidence:
classification coverage_quality_gate_failure, confidence 0.93, first
failed gate coverage (previously undetermined).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
