Policies

Wakeplane’s behavior at the edges — missed runs, concurrent execution, and failures — is controlled by per-schedule policies.

Retry policy

Controls how failed runs are retried.

{
  "retry_policy": {
    "max_attempts": 3,
    "backoff": "exponential",
    "initial_delay": "30s",
    "max_delay": "10m"
  }
}

Field	Default	Meaning
max_attempts	1	Total attempts including initial. 1 means no retry.
backoff	"fixed"	"fixed" or "exponential"
initial_delay	"1m"	Wait before first retry
max_delay	"1h"	Cap for exponential backoff

Runs that exhaust the retry budget transition to dead_letter.

Overlap policy

Controls what happens when a new occurrence fires while the previous run is still active.

Policy	Behavior
allow	Both runs proceed concurrently (default)
skip	New occurrence is skipped if prior is still running
replace	Prior run is cancelled cooperatively; new run proceeds

replace is cooperative. The executor must respect ctx.Done(). Non-cooperative executors are not forcibly terminated — they are handled by the timeout boundary.

Misfire policy

Controls what happens when the scheduler restarts and finds occurrences that were missed during downtime.

Policy	Behavior
run_latest	Run the most recent missed occurrence only (default)
run_all	Run all missed occurrences
skip	Skip all missed occurrences

Timeout

Per-run execution timeout. If the executor does not complete within this window, the run is transitioned to failed and retry/dead-letter policy applies.

{
  "timeout": "30m"
}

See also

• Concepts: Run Ledger — run states and transitions
• Concepts: Executors — non-cooperative executor behavior