Hello @JAnstee_WMF! There was some recent work we did that we hadn't released yet, but that shouldn't have been the cause of your issue. At any rate, I've updated production and I tested a few events and all seems to be working. I see there are no pending jobs in the queue, so your event shouldn't be in the "Updating" state anymore. Could you try updating it again? If you're still having problems, could you share a link to the event you're talking about?

I found this event created by you: https://eventmetrics.wmflabs.org/programs/82/events/276 It appears to have successfully finished processing just minutes after you filed this bug. It may be that it simply took quite a while to finish because you're targeting all Wikipedias. I also see some very prolific users were added as participants, which will also slow it down (more edits to scan).

That said, the update to Event Metrics we just pushed puts us on the new database cluster, which should be faster. Out of curiosity, I triggered another update for your event, just to see how long it would take now. It's still processing at the time of writing, but it started at 2021-04-02 19:44:14 (UTC). Hopefully the run time is much lower than it was before. My guess is fetching pageviews is the slowest part, since in your case it has to look up almost 50,000 individual pages.

Finished at 2021-04-02 20:10 (UTC), so it took about 16 minutes to process. That is on the long end of the spectrum, but hopefully still faster than what you had to go through earlier today. Anyways, I will wait for your confirmation before considering this resolved.

Thank you, @MusikAnimal - Good to know the why - It seems that definitely explains why this and recent attempts by Dumisani have had issues. This was my testing attempt to use a much smaller group. Thanks for your help and the improvement! I wonder, is there a way to filter the metrics queried?

@JAnstee_WMF You cannot select specific metrics to be queried, if that's what you mean. Did Dumisani's tests ever finish? I see no pending jobs in the job queue. If the amount of data is simply too much for the system, you will eventually get an error "This process took too long and was stopped". Did they run into that? If so please share a link to the event(s) and I'll see if there's anything we can do to help. Narrowing down the wikis which are queried will most certainly speed things up, if you know which specific wikis you want to query.