Quickstart.
Six short steps: sign in, connect a database, pick tables, add a destination, name the pipeline, and change a row. Five minutes if your database already has logical replication on; a few more if it needs a setting first.
Sign in
Go to app.waltail.com/get-started and enter your email. You will get a six-digit code; type it in. There is no password and no credit card. A workspace is created for you the first time you sign in.
Connect your database
Paste a connection string, for example
postgres://waltail:[email protected]:5432/app, and press Test.Waltail runs six checks: can it reach the host, do the credentials work, is the version 12 or newer, is
wal_levelset tological, may the user replicate, and is a replication slot free. Each check shows a pass or a fail.If a check fails, the console shows what to change — including the SQL to run — and you press Test again. If your database needs a setting turned on first, follow the guide for your host: RDS and Aurora, Supabase, Neon, Cloud SQL, Azure, Railway or your own server.
Pick tables
Tick the tables you want to stream. A table needs a primary key — or
REPLICA IDENTITY FULLor an index identity — for updates and deletes to replicate. The console marks any table that does not qualify and tells you why. You can change the set later.Add a destination
Choose one of the 19 destinations and fill in its fields — a queue URL, a broker address, a webhook URL, and so on. Press Test: Waltail checks it can actually reach and write to the destination before anything is saved.
You can add more than one destination to the same pipeline. Each has its own batch size and rate limit, which you can leave at the defaults.
Name the pipeline and create it
Give the pipeline a name. Waltail now creates a publication for your tables and a logical replication slot on your database — both named
pgcdc_<pipeline-name>— and starts streaming. Nothing else is created on your side.Change a row
Insert or update a row in one of the tables you picked. The console's first-event screen shows it being captured and then delivered. From here on, every committed change follows within about a second.
Next
- Have rows that existed before the pipeline? Run a backfill to deliver them through the same destinations.
- See the message format your consumer will receive.
- Read how ordering, retries and the undelivered queue work.
- Automate it with the management API.