This looks super promising, really like the deep PostgreSQL integration and the effort toward durable execution.

One aspect I’d be curious to hear more about (and might be worth expanding on in docs or future posts) is how hatchet holds up operationally in production. For example, what does a typical alerting setup look like for common failure modes? And since the system relies on partitioned tables and tuned schemas, how do you approach migrations or schema changes without downtime?

A lot of open-source job orchestration systems shine at the core execution model but fall short when it comes to observability and smooth day-2 operations. If Hatchet nails that too, it’s a huge win.