Troubleshooting

Connection failed when creating a project

The message shows what the database said rather than a generic failure, so read it before changing anything.

Authentication failed, wrong user or password.

Could not connect / timed out, usually a firewall or an unreachable host. Hosted databases often require allowlisting; check whether yours restricts inbound connections.

Database does not exist, the name after the last / in your connection string is wrong.

SSL required, set SSL mode to require.

We also refuse connections to internal and loopback addresses. If you are pointing at something on your own machine, SeedQL cannot reach it: it runs on our servers, not in your browser. Use a reachable host, or generate and download instead of pushing.

Schema is empty after connecting

You are probably on the wrong schema. Default is public; if your tables live elsewhere, set the schema name in project settings.

Push says pushed but rows are missing

Compare rows pushed with rows requested. If they differ, some rows were rejected, and the reasons are listed under the summary.

The usual cause is a unique constraint. Generated values collided with rows already present. Either truncate first, or generate fewer rows into a table with a small unique space.

Push wrote nothing

Pushed to database will say no. Every row hit a constraint. Same causes as above, at full severity: most often a primary key collision with existing data.

A table was skipped

Check it is ticked in table configuration, and that it exists in the schema SeedQL read. If you added it after connecting, re-introspect from project settings so SeedQL sees it.

Generation is slower than expected

Row count drives most of it, but foreign key depth matters too: deeply nested relationships take longer because parents must be complete before children start.

Above roughly 100,000 rows the push runs in the background. That is normal.

A scheduled snapshot says skipped

Not a failure. Most often the project has no schema captured yet, so there is nothing to snapshot. Capture one manually and the next run will succeed.

If the reason mentions your plan, scheduling is a Business feature and the plan lapsed. The schedule resumes on upgrade.

Too many requests

You hit the rate limit. The message says how long to wait, usually under a minute. If you see it during normal interactive use rather than automation, tell us, because that is not intended.

Something is still wrong

Get in touch. Include the project name and roughly when it happened; that is usually enough to find it in the logs. Bug reports go straight to the person who wrote the code.