Skip to content

feat(ops): install the self-healing host health monitor #53

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We鈥檒l occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
danielroberts20 merged 1 commit into from
Oct 7, 2026
Merged

feat(ops): install the self-healing host health monitor #53

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 6 additions & 0 deletions docs/PI_INFRASTRUCTURE_AUDIT.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -717,3 +717,9 @@ I should have checked the real data before reporting it; I inferred from run sta
**Installed 2026-10-07 ~03:37 UTC by the user and switched live immediately** (`--dry-run` removed from the user unit; timer every 30 min). The first live run's log shows both sessions seen twice; the Constellation session (idle since its last transcript write at 00:53 UTC) becomes eligible at 03:53 UTC and will be stopped on the next run after that (~04:07 UTC), with one Pushcut.

**Live behaviour (2026-10-07):** the first live run stopped the idle Constellation session at 04:07 UTC (SIGTERM was enough; the process was gone and a Pushcut was sent). Three hours later it also stopped the conversation it was being developed in, after 3 h without a transcript write while the user was away; the app resumed it from its transcript on the next message. That is why the window was raised to 6 hours.

---

## 25. Health monitor installed (2026-10-07)

`scripts/check_system_health.py` was replaced by the reviewed draft (`docs/drafts/check_system_health.py.proposed`, now removed) after about nine hours of every container reporting healthy. It was installed with an atomic rename a few minutes before a cron tick. The first real run (09:15 UTC, root's cron) created its state file, ran every new check (mounts, Docker health, restart counts, per-container OOM, systemd services, backup age, Prefect DB size, swap-in rate) and raised no alerts. The permanently-tripped "swap > 60 %" alert is gone; swap now alerts above 90 % or on sustained swap-in. Restart-on-unhealthy is rate-limited to 3 per 6 hours per container. Roll back with `git checkout <previous commit> -- scripts/check_system_health.py`.
Loading
Loading