clj-commons

Darin Douglass 2026-01-30T23:11:24.639019Z

Is there a way to make durable-queue preserve completed tasks as a means of replayability/disaster recovery? I see :complete? helps classify tasks on startup but on put! it looks like queues prune slabs and doesn't honor that opt

Darin Douglass 2026-01-30T23:23:00.129649Z

Though it looks like even with :complete? if you're actively completing tasks they’ll be cleaned up regardless since status is checked before that fn. 🤔 So it seems like “no, not out of the box”

seancorfield 2026-01-30T23:31:26.717069Z

If completed tasks are not pruned, the queues/slabs will grow indefinitely which seems... dangerous?

Darin Douglass 2026-01-30T23:38:30.695569Z

In my use-case, it'd be gated to 1 months worth of tasks (<100 GB of data)

seancorfield 2026-01-30T23:50:07.251159Z

Feel free to create a GH issue with a repro case, and an explanation of what you'd like to happen, and I can take a look. The remaining open issues on durable-queue are both hard and long-standing -- esp. the thread safety issues -- so it may not be easily solvable.

Darin Douglass 2026-01-31T00:22:35.206839Z

Ok will do. 👍

Darin Douglass 2026-02-02T16:46:15.698969Z

https://github.com/clj-commons/durable-queue/issues/31 Here's the feature request. Let me know if you want any more context

👍🏻 1