I’m pleased to announce the release of JobRunr v6.2.0 (which is now available via Maven Central) and JobRunr Pro v6.2.0 which is available for customers with a subscription.
This release includes an important bugfix for people running JobRunr on Windows machines that was reported by the community and a Pro customer. Sometimes, no jobs were processed on Windows due to a deadlock when querying CPU information. This has now been solved. (Edit 2023-05-25: I was never able to reproduce this myself but the issue has also been seen on CentOS, hence the warning on the homepage to upgrade to JobRunr 6.2 asap.)
On top of that, we did some profiling and made JobRunr a bit faster. Last but not least, we upgraded our dependencies including Spring Boot 3 and we now also support Quarkus 3.0. Support for Quarkus 2.0 and Kotlin 1.6 has dropped as we want to stay up-to-date with the latest technologies.
For JobRunr Pro customers, we added a very cool and much requested feature: dynamic queues or load-balancing!
This means your application can now support multiple tenants concurrently. Even in a scenario where one tenant, say Tenant-A, generates millions of jobs while others only create a few, these other tenants won’t be left waiting until all of Tenant-A’s jobs are completed. Instead, JobRunr utilizes a round-robin load-balancing strategy to ensure each tenant receives balanced job processing, facilitating fairness in processing their jobs.
Read all about dynamic queues in the JobRunr Pro documentation!
The hard part of building an AI agent is not the model. It is everything that has to happen reliably after the model decides what to do. Here is why background jobs are the missing piece most agent frameworks ignore.
Ever found yourself wondering whether you need a message queue or a job scheduler? It’s a common crossroads for developers. This post explains the difference.
