How can we help?

The job's lease expires after its timeout (30 seconds by default) and Halyard puts it back on the queue as a new attempt. Make handlers idempotent so a retried attempt is safe; the Idempotent jobs guide shows how.

Create a live key under Settings → API keys in the dashboard and swap it into your production environment. Test and live data are fully separate, so queues, schedules and webhook endpoints have to be created again in live mode, or promoted with halyard deploy.

Halyard guarantees at-least-once delivery. A job is written to three zones before the API answers and is only marked done when your worker acknowledges it. Pass an Idempotency-Key when you create jobs to avoid duplicates from client retries.

Each job you create counts once, however many attempts it needs. Retries, scheduled runs of the same schedule and event deliveries to webhooks do not count again. The free plan includes 10,000 jobs a month.

Projects run in us-east, eu-west or ap-south. Choose the region when you create a project; job payloads and logs never leave it. Business and Enterprise plans can add a second region for failover.

30 days on Free and Pro, 90 days on Business and up to 400 days on Enterprise. You can delete a job's payload early with DELETE /v3/jobs/{id}/payload, for example to honour a data deletion request.

Endpoints that fail every delivery for 72 hours are disabled automatically and the project owners get an email. Fix the endpoint, then re-enable it in the dashboard; deliveries from the last 7 days can be replayed.

Yes. Pro projects can request up to 500 requests per second from the dashboard; Business and Enterprise limits are set per contract. Include your project ID and expected peak traffic in the request.

A data processing agreement is available on every paid plan from Settings → Legal. SAML SSO and SCIM provisioning are included in Business and Enterprise.