Jobs

A Job runs your code once and finishes. Use it for imports, reports, backfills, data migrations, file processing and other batch work. Unlike a Function, nobody waits for a response: you start a run, and Coderunner records whether it succeeded or failed.

To run a Job on a timetable, use a Schedule instead. If your code needs to keep running, use a Service.

Write the handler

A Job uses the same files and handler as a Function, but it doesn't need to return anything:

const handler = async (context, payload) => {
  console.log("Starting import", payload);
  const count = await importCustomers(payload.source);
  console.log(`Imported ${count} customers`);
};

module.exports = { handler };

payload is the JSON you start the run with; it can be empty. In .NET, Execute returns Task; in Rust, handler returns (). See the language references.

Deploy it

gf deploy --name customer-import --workload-type job

Deploying doesn't run the Job. It makes it ready to run.

Start a run

A Job also has a Deployment URL, mainly so you can start test runs; call it as you would a Function.

Success and failure

A run finishes as Succeeded or Failed:

• If an error escapes your handler, the run is Failed, and its error details are kept with the execution.
• If you catch an error yourself, the run finishes as Succeeded. That's your choice: handle it, retry, or log it and move on.

Anything your handler prints appears in the Process Log, so log enough to understand a run afterwards. Executions, with their status and error details, are also listed by the API: GET https://platform.groundfloor.cloud/v1/workspaces/{workspace_id}/executions.

Next steps

• Schedules — run this Job automatically.
• Configuration and secrets — settings and secrets your Job can read.
• Logs and troubleshooting — find out why a run failed.