Schedules and events

Audience: teams starting workflows from time-based or event-driven triggers.
Works with: schedules work in OSS and Orkes; event infrastructure depends on server configuration.

Use the typed SchedulerClient for schedule metadata and the workflow EVENT task or event handlers for asynchronous events. The Java API map links each client and its canonical Javadocs.

Schedule safely

Use a stable correlation identifier derived from the time window so a retry or double delivery cannot create duplicate business work. Keep the scheduled workflow idempotent and make its input small: pass object references, not large files.

Fragment — create the client from OrkesClients; the API map links the complete signature reference.

StartWorkflowRequest start = new StartWorkflowRequest()
        .withName("billing")
        .withVersion(2)
        .withCorrelationId("billing-2026-07-19")
        .withInput(Map.of("billingDate", "2026-07-19"));

SaveScheduleRequest schedule = new SaveScheduleRequest()
        .name("daily-billing")
        .cronExpression("0 0 2 * * ?")
        .zoneId("UTC")
        .startWorkflowRequest(start);

schedulerClient.saveSchedule(schedule);

Expected result: the schedule is visible through the server API/UI and produces a workflow execution at the configured time.

Event-driven work

Use an EVENT task to publish from a workflow, or configure an event handler to start a workflow. Put event payload validation at the workflow boundary and include an idempotency key from the producer.

Next: workflow lifecycle, reliability, and debugging.