You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The pipeline auto-rolls back on health-check failure, so production stays on the previous release. Check the failed run log for the root cause — common causes: a missing required env key (see REQUIRED_ENV_KEYS in deploy.yml) or a startup config error in the new binary.
Keep this issue open until the root cause is fixed and a deploy succeeds.
The deploy fails at the Upload binary to VPS step (appleboy/scp-action), not in the code.
deploy.yml SCP step + CONTABO_ secrets are unchanged* since the last successful deploy (2026-11-05 01:34).
VPS is up: ports 22/80/443 open; https://bigbase.click/api/monitoring/health returns {"status":"ok"} (old release v2.93.0 still serving).
Two attempts failed identically (workflow_run + workflow_dispatch redeploy).
This points at VPS-side SSH state changed by the earlier rollback (e.g. authorized_keys restored from a snapshot, fail2ban banning GitHub runner egress, or /tmp disk full), not a repo regression.
Suggested check on the VPS (or via Contabo console): df -h /tmp, fail2ban-client status sshd, and confirm the deploy key is still in ~/.ssh/authorized_keys. Then re-trigger the Deploy workflow for run 42305803617.
Deploy succeeded — v2.94.1 is live on production (footer confirms). Root cause was the VPS root filesystem at 100% (accumulated site build artifacts in /opt/bigbase/data/builds + writable). Freed space by clearing caches and deleting stale build/backup/release leftovers (kept the 9 running deployments). Re-triggered the Deploy workflow for run 42315255890 — all steps green.
Activity
danielvm-git commentedon Nov 6, 2026
Diagnosis (2026-11-06)
The deploy fails at the Upload binary to VPS step (
appleboy/scp-action), not in the code.https://bigbase.click/api/monitoring/healthreturns{"status":"ok"}(old release v2.93.0 still serving).This points at VPS-side SSH state changed by the earlier rollback (e.g.
authorized_keysrestored from a snapshot, fail2ban banning GitHub runner egress, or/tmpdisk full), not a repo regression.Suggested check on the VPS (or via Contabo console):
df -h /tmp,fail2ban-client status sshd, and confirm the deploy key is still in~/.ssh/authorized_keys. Then re-trigger the Deploy workflow for run42305803617.github-actions commentedon Nov 7, 2026
Another deploy failure: https://github.com/danielvm-git/bigbase/actions/runs/42316028080
danielvm-git commentedon Nov 7, 2026
Deploy succeeded — v2.94.1 is live on production (footer confirms). Root cause was the VPS root filesystem at 100% (accumulated site build artifacts in /opt/bigbase/data/builds + writable). Freed space by clearing caches and deleting stale build/backup/release leftovers (kept the 9 running deployments). Re-triggered the Deploy workflow for run 42315255890 — all steps green.