Operations & observability¶
How adunopay services are monitored and where to find logs when something goes wrong. For where each service runs, see Hosting & infrastructure.
Some details are unconfirmed
Entries marked (to confirm) are best-guess and need verifying by someone with infra access. Correct them in place and drop the marker once checked.
Monitoring¶
Where we watch service health and uptime, and how the team gets alerted.
Scope: operational health only
This covers whether services are up and healthy. Monitoring for suspicious activity (payment fraud / AML and network-security threats) is a separate risk/compliance concern — see Suspicious-activity monitoring.
Not yet documented — to confirm
It is not yet known what monitoring is in place. To fill in:
- Is there any uptime monitoring on the public endpoints (portals, API), and if so, what tool? (An external service like UptimeRobot, Better Stack, Pingdom, or a self-hosted check — to be determined.)
- Which endpoints/services are checked, and at what interval?
- Where do alerts go (email, Slack, on-call)?
- Is there any application/infra metrics monitoring (CPU, memory, disk, DB) on the hostpoint server, or only external uptime checks?
- Are the Vercel frontends monitored separately (Vercel Analytics / deployment alerts)?
Logging¶
Where to find logs for each tier.
| Tier | Where logs live | Notes |
|---|---|---|
| Backend / API (Node.js + nginx) | On the hostpoint.ch server (to confirm) | e.g. journalctl / nginx access & error logs / app log files — path TBD. |
| Frontends (4 portals) | Vercel dashboard → project → Logs (to confirm) | Build and runtime/function logs per deployment. |
| Database (Postgres 16) | On the DB host (to confirm) | Postgres logs; location depends on where the DB runs. |
To confirm
- Exact log locations/paths on the backend server.
- How long are logs retained, and are they rotated?
- Is there any centralized log aggregation (e.g. a hosted log service), or are logs only on each host?
- How does someone get access to read them?