Monitoring: /health for uptime checks, structured error logging, API errors without internal details #90

anas · 2026-10-09T17:30:47Z

anas commented

2026-10-09 17:30:47 +00:00

Closes #82, and the API side of #81. The server side (log rotation, /health through the preview gate, scripts/errors.sh, how to set up an uptime monitor) is in the private repo divan-deploy.

/health (site): asks the API, which asks the database (3 s timeout).

200 {"ok":true,"site":"ok","api":"ok","db":"ok","ms":…} when all three answer;
503 naming the part that is down otherwise;
never cached.

An uptime monitor (UptimeRobot, Better Stack…) checks it and alerts the owner.

API

GET /api/health checks the database.
One error handler: 4xx keep their message. Server errors answer {"error":"سرور میں خرابی۔ تھوڑی دیر بعد کوشش کریں۔"}, never the internal message, and are logged with the method, address and the whole error.

Site: any error while making a page is logged as one JSON line (time, method, path, message, first lines of the stack). The reader sees the Urdu error page from #88.

Tested with test servers:

healthy: 200;
database unreachable: 503 db: down;
API unreachable: 503 api: down;
an API request with the database down: 500 with the Urdu message, logged with method, address and message;
bad JSON: 400 with its message;
a page with the API down: a JSON error line with the path.
npm test passes; one test is skipped by design (it needs the data export).

🤖 Generated with Claude Code

Closes #82, and the API side of #81. The server side (log rotation, `/health` through the preview gate, `scripts/errors.sh`, how to set up an uptime monitor) is in the private repo divan-deploy. **`/health`** (site): asks the API, which asks the database (3 s timeout). - `200 {"ok":true,"site":"ok","api":"ok","db":"ok","ms":…}` when all three answer; - `503` naming the part that is down otherwise; - never cached. An uptime monitor (UptimeRobot, Better Stack…) checks it and alerts the owner. **API** - `GET /api/health` checks the database. - One error handler: 4xx keep their message. Server errors answer `{"error":"سرور میں خرابی۔ تھوڑی دیر بعد کوشش کریں۔"}`, never the internal message, and are logged with the method, address and the whole error. **Site:** any error while making a page is logged as one JSON line (time, method, path, message, first lines of the stack). The reader sees the Urdu error page from #88. **Tested** with test servers: - healthy: 200; - database unreachable: 503 `db: down`; - API unreachable: 503 `api: down`; - an API request with the database down: 500 with the Urdu message, logged with method, address and message; - bad JSON: 400 with its message; - a page with the API down: a JSON error line with the path. - `npm test` passes; one test is skipped by design (it needs the data export). 🤖 Generated with [Claude Code](https://claude.com/claude-code)

anas added 1 commit 2026-10-09 17:30:48 +00:00

Monitoring: /health (site, API, database) for an uptime monitor; errors logged as one JSON line each by the API and the site; the API answers errors without internal details 0511f7d2ee

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

anas merged commit 415459d55f into main

2026-10-09 19:49:08 +00:00

anas referenced this issue from a commit

2026-10-09 19:49:09 +00:00

Merge pull request 'Monitoring: /health for uptime checks, structured error logging, API errors without internal details' (#90) from feature/monitoring into main

anas referenced this pull request

2026-10-09 22:34:47 +00:00

Launch: Urdu 404 and error pages #81

Sign in to join this conversation.

No reviewers

No Label

No Milestone

No project

No Assignees

1 Participants

Notifications

Due Date

The due date is invalid or out of range. Please use the format 'yyyy-mm-dd'.

No due date set.

Dependencies

No dependencies set.

Reference: anas/divan#90