Skip to main content
GET
Tenant-wide event stream — receive status events for every job. Same SSE shape as /jobs/{id}/events but unscoped from a single job. Useful for tenant dashboards that want to render an in-flight panel without fanning out one connection per job. Events are filtered server-side: only events whose id belongs to the calling tenant are surfaced. The first lookup for a previously unseen job hits the DB; subsequent events for the same job are served from an in-stream cache. Behaviour:
  • 15s keepalive comments to survive proxies.
  • 30-min lifetime cap; client should reconnect after event: timeout.
  • No initial snapshot (the per-tenant set could be huge); use GET /jobs?status=running if you need a starting state.

​
Response

Returns any.
The spec this page is generated from declares this response as an untyped object — the server returns JSON, but its shape is not published. Read a live response to see the fields; treat any field you find as unversioned until the spec names it.

​
Status codes