When to use something else
Threadmill is a job engine that lives inside your application. It is the wrong tool for some problems, and this page says which.
| If you need | Consider | Why |
|---|---|---|
| Workflows that run for weeks, wait for signals, or span services in several languages | Temporal or Restate | They are workflow platforms with their own server and replay model. Threadmill runs jobs, not long-lived orchestrations |
| Step-level checkpoints inside a single function, on PostgreSQL | DBOS Transact or JobRunr's runStepOnce | Threadmill retries a whole job; your handler must be idempotent |
| Atomic batches or rate limiting out of the box | JobRunr Pro | Threadmill does not have these yet |
| Quarkus or Micronaut integration | JobRunr | Threadmill ships a Spring Boot starter; the core is framework-free but other adapters are not packaged |
| The smallest possible scheduler | db-scheduler | One table and one dependency |
| One scheduled method that must not run on two nodes at once | Spring @Scheduled with ShedLock | A lock is enough; you do not need a job engine |
| Chunked batch processing of large data sets | Spring Batch | Often used together with a scheduler such as Threadmill |
| A Java or database version outside the compatibility table | JobRunr, db-scheduler or Quartz | They support older runtimes |
Checked on . Sources: Temporal documentation, Restate documentation, DBOS documentation, JobRunr documentation, JobRunr Pro documentation, db-scheduler README, ShedLock README, Spring Batch, Quartz documentation.
What Threadmill is for
Durable background work inside a Spring Boot service: sending email, calling other systems, generating documents, syncing data, running scheduled maintenance. Work that must survive a restart, retry on failure, respect concurrency limits and be visible to the people on call.