When a workspace is paused

Durablex Cloud pauses a workspace that reaches its plan's usage limit. Pausing gates the workspace against new work: the engine refuses to start new top-level runs, while everything already in flight runs to completion and reads stay available. It is a binary switch - new work stops, in-flight work drains.

What suspension does

When a workspace is paused:

• New event-triggered runs are refused. POST /events still returns 202 and still wakes any in-flight runs waiting on the event, but it starts no new run and the response carries "suspended": true with an empty triggered list.
• Cron schedules skip. A due tick advances to its next fire time and is skipped (logged, not an error); it is not backfilled. When the namespace is resumed, the next tick fires normally.
• Replay is refused with 403 Forbidden.
• Debounce and batch buffers are held. A coalescing or batch buffer that comes due while the namespace is suspended is not fired and not dropped - it waits in the store and drains when the namespace is resumed, so suspension never starts a run from buffered work either.
• In-flight runs finish. Child runs, on-failure handlers, queued runs, and runner callbacks all proceed - pausing blocks new runs, not the execution of runs already started.
• Reads are unaffected. Listing runs, streaming, and the console keep working, so you can still see your workspace's state.

How pausing works on Cloud

Pausing is automatic. When your workspace reaches its plan's included usage for the period, it pauses; it resumes at the start of the next period, or immediately when you upgrade to a plan with more headroom. The console shows a paused banner, and the Usage view tracks how much of each limit you've consumed so you can see a pause coming.