Confluence Server Down Frequently (twice in a month)

[Atlassian B2X]

My confluence down once in every two weeks, Showing some caches stuck:

13-Aug-2021 06:33:16.624 WARNING [Catalina-utility-3] org.apache.catalina.valves.StuckThreadDetectionValve.notifyStuckThreadDetected Thread [http-nio-8090-exec-240] (id=[23293]) has been active for [68,479] milliseconds (since [8/13/21, 6:32 AM]) to serve the same request for [https://wiki.b2xcare.com/plugins/servlet/eb/search] and may be stuck (configured threshold for this StuckThreadDetectionValve is [60] seconds). There is/are [47] thread(s) in total that are monitored by this Valve and may be stuck.

[Radek Dostál]

This wiki.b2xcare.com site - how does it come in play in your environment, is it an application link? We've for instance once observed that when another application over an applink goes down, our Confluence couldn't breathe either, I don't remember if there were similar stuck threads or not at the time but I'd imagine so.

No other guesses apart from that without knowing how that url is used. I don't otherwise recognize the "/plugins/servlet/eb/search" path, and it does not open in our Confluences, so if this is Confluence, then I will assume it's a plugin path - maybe a problem with the plugin itself causing the threads to become stuck, that can happen as well.

[Atlassian B2X]

Yes Radek,

https://wiki.b2xcare.com/ this is URL of my confluence.

[Radek Dostál]

In that case, if you're not seeing stuck threads for other things other than the /plugins/servlet/eb/search, my guess would be the plugin implementing this endpoint is causing the stuck threads. To me this sounds like a good starting point to add more logging to the plugin to help pin point the cause, and also to take thread dumps when the stuck threads are occurring to find out at which point those threads became stuck and send them over to the developer.

I'm still focused on that this endpoint is not an out of box feature so it's added by something and it's very probable that that something is causing it, possibly some iffy implementation, something doesn't run as should and times out. Thread dumps should reveal more.