Short answer: there isn’t a deployment-level “max runtime” knob in Prefect 3.x. Instead, set time limits at the flow/task level or use an Automation to cancel long-running runs. Options: - Flow/task timeouts (most common) - Define a hard cap with

timeout_seconds

on the flow and/or tasks:

from prefect import flow, task

  @task(timeout_seconds=120)  # 2-minute cap per task attempt
  def work():
      ...

  @flow(timeout_seconds=600)  # 10-minute cap for the whole flow run
  def my_flow():
      work()

- You can override at deploy time:

my_flow.with_options(timeout_seconds=900).deploy(
      name="my-deployment",
      work_pool_name="default"
  )

- Note: timeouts mark the run as failed and attempt to cancel; truly preempting blocking sync code in threads isn’t always possible. Prefer async I/O or

ProcessPoolTaskRunner

for stronger preemption. Details: Task/flow timeout behavior - Automations to cancel long runs (Cloud) - Create an Automation that cancels any flow run that stays Running longer than X minutes:

from datetime import timedelta
  from prefect.automations import Automation
  from prefect.events.schemas.automations import EventTrigger
  from prefect.events.actions import CancelFlowRun

  Automation(
      name="Cancel long-running flows",
      trigger=EventTrigger(
          after={"prefect.flow-run.Running"},
          expect={"prefect.flow-run.Completed"},
          posture="Proactive",
          within=timedelta(minutes=30),
          for_each={"prefect.resource.id"},
      ),
      actions=[CancelFlowRun()],
  ).create()

- Metric-based triggers (e.g., “duration > N seconds”) are also available in Cloud. - Docs: Create automations, Event triggers, Cancel workflows If you can share whether you’re on Prefect Cloud or OSS, and whether you want a strict hard-stop at N minutes or just a cancellation after N minutes, I can suggest the best-fit pattern and example for your setup.