72002: LiveSync database connector is unhealthy

A LiveSync database connector could not read from the table or collection it is configured to use, because that table or collection is missing, renamed, or no longer accessible to the connector.

What you should do

The connector runs a periodic health check on the database it reads from, and this error means that check failed. What to confirm depends on which connector you are using.

For a Postgres connector, confirm the outbox and nodes tables it uses still exist, are named correctly in the connector configuration, and that the connector's database user still has access to them.

For a MongoDB connector, confirm the watched collection still exists, is named correctly in the connector configuration, and that the connector's database user still has the find and changeStream permissions on it.

Once the underlying problem is resolved, the connector recovers on its next health check; no manual restart is needed.

Why it happens

A LiveSync database connector continuously reads changes from a table or collection and publishes them to Ably. This error means the connector could not open or query that table or collection, because it is missing, has been renamed, or its permissions have changed. A database the connector cannot connect to at all is reported as 72003 instead, and an individual change that cannot be mapped to a channel as 72004.

What you'll see

The error is reported with code 72002 and HTTP status 500. The message names the affected table and takes the form unable to connect to table: <name>: table is unhealthy: <detail>, or unable to connect to tables: <detail> when the connector cannot open its tables at startup.