Fix health check: accept HTTP 503 (stalled) as healthy
Some checks are pending
CI / lint-and-test (push) Waiting to run
Some checks are pending
CI / lint-and-test (push) Waiting to run
Pipeline /health returns 503 when idle/stalled, which is a valid running state. Also increase post-restart wait from 15s to 30s for pipeline HTTP server initialization. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This commit is contained in:
parent
fe996c3299
commit
25062cf130
1 changed files with 5 additions and 4 deletions
|
|
@ -96,7 +96,7 @@ fi
|
||||||
if [ -n "$RESTART" ]; then
|
if [ -n "$RESTART" ]; then
|
||||||
log "Restarting:$RESTART"
|
log "Restarting:$RESTART"
|
||||||
sudo systemctl restart $RESTART
|
sudo systemctl restart $RESTART
|
||||||
sleep 15
|
sleep 30
|
||||||
|
|
||||||
FAIL=0
|
FAIL=0
|
||||||
for svc in $RESTART; do
|
for svc in $RESTART; do
|
||||||
|
|
@ -110,10 +110,11 @@ if [ -n "$RESTART" ]; then
|
||||||
done
|
done
|
||||||
|
|
||||||
if echo "$RESTART" | grep -q "teleo-pipeline"; then
|
if echo "$RESTART" | grep -q "teleo-pipeline"; then
|
||||||
if curl -sf --connect-timeout 3 http://localhost:8080/health > /dev/null 2>&1; then
|
HEALTH_CODE=$(curl -s -o /dev/null -w '%{http_code}' --connect-timeout 3 http://localhost:8080/health 2>/dev/null || echo "000")
|
||||||
log "pipeline health: OK"
|
if [ "$HEALTH_CODE" = "200" ] || [ "$HEALTH_CODE" = "503" ]; then
|
||||||
|
log "pipeline health: OK (HTTP $HEALTH_CODE)"
|
||||||
else
|
else
|
||||||
log "WARNING: pipeline health check failed"
|
log "WARNING: pipeline health check failed (HTTP $HEALTH_CODE)"
|
||||||
FAIL=1
|
FAIL=1
|
||||||
fi
|
fi
|
||||||
fi
|
fi
|
||||||
|
|
|
||||||
Loading…
Reference in a new issue