Self-host Traceway in 90 seconds

Your data, your metal. The self-host build is the same software Cloud runs — same map, every stop — and standing it up is one compose file.

The whole yard in one command

$ git clone github.com/tracewayapp/traceway
$ cd traceway && docker compose up -d
# clickhouse · postgres · collector
✓ dashboard at http://localhost:3000

Three services come up:

• ClickHouse stores the events — a million events a day is roughly 2 GB a month at rest.
• Postgres holds the metadata: projects, alerts, users.
• The collector speaks OTLP. Point your existing OpenTelemetry SDKs at it and you're in — no vendor SDK, no agent to babysit.

Zero config is the config

There is no step two. No sampling rules to tune before you see data, no retention policy wizard, no license key. The defaults are the ones Cloud runs.

When you'd rather not staff the yard

Cloud is the same line with the maintenance stops removed — we run the ClickHouse, the upgrades, and the 3am disk alerts. The fare is two meters (exceptions and ingested data), agreed up front. And the door swings both ways: Cloud teams can export and come back to self-host any time.