Scheduler

The Scheduler runs an agent on its own at set times, without anyone sending it a message. Each time the schedule fires, the agent receives the Scheduler Input you configured and runs as if a user had sent it. Failed runs are retried automatically, and every run is recorded in the Executions history. Use it for recurring work such as daily summaries, periodic data checks, or weekly reports.

​

Creating a schedule

Open the agent in Agent Builder and open its scheduler settings. The Scheduler Configuration panel opens.

1. Pick how often the agent runs using one of the timing tabs described below.
2. Set Max Retries and Retry Delay to control what happens when a run fails.
3. Enter the Scheduler Input: the message the agent receives on every run, for example Summarize yesterday's support tickets and flag anything urgent.
4. Select Create Scheduler.

The schedule is active as soon as it is created.

​

Timing options

A cron expression has five fields: minute hour day-of-month month day-of-week. Day of week runs from 0 (Sunday) to 6 (Saturday).

​

Retries

​

Monitoring runs

Select Executions in the panel header to open the run history for the agent’s schedules. Each entry shows:

• Status: whether the run succeeded or failed.
• Error: the reason a run failed, for example a request timeout.
• Time: when the run happened, so you can confirm the schedule is firing on its intended cadence.

​

Managing schedules with the API

Everything in this panel is also available over REST, including pausing, resuming, and running a schedule immediately for testing. See the Schedules API. If you need to call an arbitrary URL on a schedule rather than an agent, use HTTP Schedules.

​

Related

• Webhook Trigger: run an agent when an external system sends it an event, instead of on a timer.