feat(demo): provoke the triage tick instead of waiting a minute for it
Some checks failed
Tests / Declarative: Post Actions failed: 2, passed: 142

Both demos went quiet for up to a minute between the build turning red and
the monitor reacting, because Ariadne's tick is on cron. The scripts now run
that tick immediately over the pod's own loopback - nothing exposed outside
the cluster - and print what it saw, so the pause becomes a visible step
rather than dead air.

Falls back to silence rather than failure: if the request does not land the
scheduler still picks the build up within the minute, which is exactly the
old behaviour.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This commit is contained in:
jenkins 2026-08-07 01:44:03 -03:00
parent 8899b76d07
commit bd6c0678d8
3 changed files with 26 additions and 0 deletions

View File

@ -175,6 +175,10 @@ PY
note "HTTP $(jenkins_post "/job/$CODE_JOB/build")"
note "result: $(wait_for_build "$CODE_JOB" "$next_num")"
say "Asking Ariadne to look now rather than on its next minute"
note "\$ POST /api/internal/hermes/autotriage/run"
note "$(poke_ariadne)"
say "Ariadne collects evidence and asks Hermes for a minimal patch"
note "Hermes returns an anchored patch as data; Ariadne validates path, size,"
note "changed lines, and that the anchor is unique, then pushes hermes-repair/$next_num"

View File

@ -92,6 +92,24 @@ wait_for_build() { # job number [max_seconds] -> prints result
printf 'TIMEOUT'
}
# Ariadne's triage tick is on cron, which cannot fire more often than once a
# minute. That minute is the largest gap between a build going red and the
# system visibly reacting, and it is pure dead air on stage. This runs the same
# tick immediately over the pod's own loopback, so nothing is exposed outside
# the cluster. The tick is idempotent - incidents dedupe on job and build
# number - so provoking it can only ever be a no-op, never a second incident.
poke_ariadne() {
kubectl -n maintenance exec deploy/ariadne -c ariadne -- python3 -c "
import json, urllib.request
req = urllib.request.Request(
'http://127.0.0.1:8080/api/internal/hermes/autotriage/run', method='POST')
body = json.load(urllib.request.urlopen(req, timeout=120))
jobs = body.get('jobs') or {}
print(body.get('status', 'ok'), '|', ', '.join(
f\"{name}={info.get('status')}\" for name, info in jobs.items()) or 'no jobs')
" 2>/dev/null || echo "tick request failed; the scheduler will pick it up within a minute"
}
ariadne_ticks() { # tail the autotriage decisions in human-readable form
kubectl -n maintenance logs deploy/ariadne -c ariadne --tail="${1:-400}" 2>/dev/null |
grep 'hermes autotriage tick' |

View File

@ -68,6 +68,10 @@ cmd_run() {
say "Waiting for the seeded build to fail"
note "result: $(wait_for_build "$FIXTURE_JOB" "$next_num")"
say "Asking Ariadne to look now rather than on its next minute"
note "\$ POST /api/internal/hermes/autotriage/run"
note "$(poke_ariadne)"
say "Ariadne detects, gathers evidence, asks Hermes, authorizes, repairs"
note "the repair is a single in-process ConfigMap patch, so watch the fixture"
for _ in $(seq 1 40); do