We have a number of tasks running on the server in various different environments. The tasks which are all running on one environment get queued, which is fine, and we can see what tasks are blocking the queued tasks (see queue_ok.jpg). But since moving to Octopus 3, we are seeing tasks getting queued on the server with nothing blocking them for several minutes (see queue_bad.jpg for an example where the task was queued for 12 mins).
I’ve checked on the server itself and the cpu and memory seem normal, not excessive. Is there anything I can do to see why these jobs are getting queued?
There are a decent number of tasks for one of our environments which are queued correctly, but I would expect other tasks for different environments to not be queued at all. Attached a pic of the task page.
Sorry for the delay, it’s a bit of a weird issue. What generally happens to unblock them? Does it just clear itself or do you have to cancel or reset or restart the service?
I might have to get you to run a query when this happens so we can figure it out. I would get you to schedule a call, but what are the chances do you think you could make this reliably happen?
I think we might have figured this out. Can you go to your Configuration -> Node page and see what the task cap is?
You are running a few tasks, and we do have the new task capping feature that is set to 5 by default. You should be able to safely increase this (I have one customer who set it to 70).
Please let me know if that helps!
Yes, that looks like it was the problem. I’ve updated the cap to 50 and the tasks for the new environment are starting immediately. Great catch!
Do you think that there could be a message in the task saying that it had been queued because of the task cap? I don’t think I would have ever worked out that that was the problem on my own.
I absolutely agree. I’ve put a task in for when this occurs to display a message to show why it was queued and how to change the setting - not that it will help you now
