golem-recurring-task-rust

Have the agent call its generated client's schedule_ method on itself at the end of each invocation, passing a Datetime set to the desired future time. This creates a durable loop where each run schedules the next one.

Use the schedule_cancelable_ variant of the generated client method, which returns a CancellationToken you can store and cancel later. Alternatively, use a boolean state flag checked at the start of the method, or run golem agent invocation cancel from the CLI with the idempotency key.

Yes, scheduled invocations are durable. If the agent crashes or restarts, the pending scheduled invocation still fires at the designated time and the agent recovers to execute it.

Track consecutive failures in agent state, then compute the delay as base interval times 2 raised to the failure count, capped at a maximum interval. Reset the counter to zero and restore the base interval on success.

Each scheduled tick appends entries to the agent's oplog, so recovery replays the full history as it grows. Enable periodic snapshotting with the snapshotting attribute so recovery starts from the latest snapshot instead of replaying every tick.