Could you please reactivate our serverless agent? ...
# dagster-plus
t
Could you please reactivate our serverless agent? It was spun down after the trial ends. Asking in case the reactivation process is not yet automated. 🙏
dagster/agent_id:7ec7f869-bb3b-4d03-a230-cc7a8a95b96c
j
I can set this up now - sorry about that!
❤️ 1
what’s your organization?
t
kovalee
j
i see it as already active
but i see a very deep run queue - i have a feeling it probably can’t keep up fast enough. let me dig a bit deeper to see if something is up.
t
Thank you. I thought it stopped taking jobs out of the queue because we ran out of trial compute minutes...
d
Hi Toky - we're looking into this now. Just confirming, is the number of runs in this deployment intentional? You're aware of the limitations for using serverless described here? https://docs.dagster.io/dagster-cloud/deployment/serverless#limitations
t
The number of runs is intentional. However, I wasn't aware of the limitations of using serverless. Thanks for the info
Is it a limitation issue instead of an end of trial one?
d
Unrelated to your trial - we're looking into this and should have things back up and running shortly
(and unrelated to serverless limits - just making sure you wanted these all to run before we got them started again)
t
ok, perfect. Thanks
j
How fast do you expect each of these to run? There are some throughput defaults in serverless and I just want to do some napkin math to see if we there’s enough throughput to eventually chew through this whole backlog.
Also, do you expect to continue enqueuing runs at this rate? Or does this represent a one time backfill?
t
This is a one time backfill.
👍 1
Each run takes less than 5s
👍 1
j
You’re back up and running now - sorry about the interruption!
We’ll keep an eye on this specifically throughout the day to make sure it continues to work through the backlog.
t
Amazing, thanks for your help
j
Just to set expectations though - it’s still going to take a long time to catch up.
t
Yes, that's fine.
FYI, I'll change the strategy and will just terminate all runs that are currently queued for now.