zwitschi/linearvoid
1
0 Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 9 Wiki Activity

feat: monitoring exporters, alerting, rate limiting, backups, incident runbook
CI / test (push) Successful in 58s
Details
CI / docker (push) Successful in 2m22s
Details
CI / release (push) Skipped
Details

Browse Source 
- Add 4 Prometheus exporters: nginx (9113), redis (9121), postgres (9187), node (9100)
- Create 8 alerting rules in monitoring/prometheus/alerts.yml
- Update prometheus.yml with all scrape targets + rule_files
- WS connection rate limiting: max 5/IP/60s (Redis INCR)
- Admin login rate limiting: max 5/IP/5min (Redis INCR)
- 3 backup scripts: daily to R2 (14d), weekly full (8w), verify with restore test
- docs/incident-runbook.md with 5-tier triage + escalation
- Update CD workflow: full prometheus config + alerts seeding
This commit is contained in:
zwitschi committed 2026-08-06 19:31:40 +02:00
1 parent b5710548b1
commit 0f81a81029
11 files changed
+683 -73
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

.github/MEMORY.md
-8
View File
@@ -1,9 +1 @@
# MEMORY.md
## 2026-08-05: Chat Message Admin + CI/CD + Release 1.0.5
**Completed**: Chat Message Admin, CI/CD, Release 1.0.5 committed + tagged + pushed.
**Bug — nested admin page**: Fixed. Root cause: page HTML files were in `frontend/admin/` (Vite source root). Vite multi-page mode injected scripts/styles into every `.html` entry. `fetch("/admin/dashboard.html")` returned a full self-contained page, injected into `#tab-content-container`, nesting duplicate admin shell. Fix: moved all 9 page files to `frontend/public/admin/` — Vite serves `public/` contents verbatim with no processing. Fetch now returns raw HTML fragment.
**Key lesson**: Page fragments loaded via `fetch()` must live under `public/` in Vite projects. Any `.html` in the project source root gets processed as a multi-page entry.