Running it
Upgrades, backups, certificates, alerts.
Two small services, so this is short. These are the things that actually go wrong.
Upgrades
Update the service before pushing new app builds.
docker compose -f infra/docker-compose.yml pull
docker compose -f infra/docker-compose.yml up -dMigrations run on start. Take a backup first.
Backups
The database is the only part that holds anything: accounts, computers, teams, history. Back it up, and restore it somewhere now and then. A backup you have never restored is a guess.
Certificates
The most likely outage by far. Every computer keeps a connection open to the coordinator, so an expired certificate looks exactly like a network failure and hits everyone at once.
What to watch
| Signal | Why |
|---|---|
| Coordinator reachable | Nothing connects without it |
| Relay reachable | Relayed sessions fail while direct ones work, which is confusing |
| Certificate expiry | See above |
| Database disk and connections | The usual |
| Relay bandwidth | The only real cost |
Capacity
The coordinator is idle nearly all the time. The relay is the one that grows, and only with sessions that cannot go direct.
Lots of relay traffic? Working IPv6 on your users' networks turns most of it back into direct. See direct and relay.
When something is wrong
- Check Settings in the dashboard.
- Check the service logs.
- Confirm the computer showing offline can reach your server.
Client-side symptoms: it will not connect.