Manuel Astudillo
Manuel Astudillo
Thanks for this FR. If I understand you correctly, this should be pretty easy to implement, at least for standard jobs, as we would just need to pick the oldest...
You can already do this using getJobs: https://api.docs.bullmq.io/classes/v5.QueueGetters.html#getJobs.getJobs-1 Just choose one job and use the asc argument appropriately so you will get the oldest in every case. For delayed jobs...
Do you have some code that reproduces the issue?
It would not be noise if you provided a minimal case that reproduces it. If I have to guess it seems like somehow you are passing giberish to the ```Queue.add```...
https://docs.bullmq.io/guide/telemetry
Closing as there is no code to reproduce it. Furthermore this issue probably does not exist anymore with the new upsertJobScheduler method.
I did try a linked list implementation, but when I ran the benchmarks it was much slower than using the old implementation. Did you notice any performance improvements?
I think the array implementation is so efficient in V8, that even an algorithmically better implementation may not give the expected results...
> ----- Average ------ > numJobs : 100 > array ms average: 0.02 > list ms average : 0 > numJobs : 1000 > array ms average: 0.32 > list...
I run the tests in https://github.com/taskforcesh/bullmq-bench but I cannot find any noticeable improvement or decrease in performance. In my M2 using old implementation 43307 jobs/sec, using new 43651 jobs/sec. But...