We had an unusual experience the other day, and I'...
# gooddata-cn
We had an unusual experience the other day, and I'm wondering if anyone has insights / thoughts to help explore the root cause. Starting at precisely 8am local time, GoodData started increasing database connections ultimately reaching approximately 20,000 simultaneous connections before we shut it down. We are still in development, so there were no active users at the time. Are there any cron jobs or other schedules that might have started this? Any related issues or situations you're aware of?
Am I correct that you are talking about connections to your data source, right? There is only one cron job cleaning cache tables, which does something only if you utilize dataSource.enableCaching and it is running only in single thread, so it should not cause this issue. @Robert Moucha any thoughts?
Correct, there are no other cron jobs except
that triggers cache purging every hour
I do not want to speculate here about a potential root cause. We have not observed such a behavior before. Please, next time this happens, generate support bundle and share it with us, so we can investigate it: https://community.gooddata.com/administration-kb-articles-49/gooddata-cn-collecting-diagnostics-support-bundle-363
Correct, it was connections to the data source. I'll work with our team on the diagnostics side and check into the cron job as well. Thanks for the notes, and I'll report back if we find anything