Spam from Google FeedFetcher requests cause CPU spike
Platform notice: Server and Data Center only. This article only applies to Atlassian products on the Server and Data Center platforms.
Support for Server* products ended on February 15th 2024. If you are running a Server product, you can visit the Atlassian Server end of support announcement to review your migration options.
*Except Fisheye and Crucible
Problem
When JIRA server is open for public, JIRA threads has reached 100%, CPU usage spike at 90% - 100% and become unaccessible.
A lot of the following appears in the atlassian-jira-http-access.log
:
"GET https://jira.atlassian.com/issues/ HTTP/1.1" - - - - "Feedfetcher-Google; (+http://www.google.com/feedfetcher.html; feed-id=xxxxxx)"
"GET https://jira.atlassian.com/issues/ HTTP/1.1" 200 xxx xxx - "Feedfetcher-Google; (+http://www.google.com/feedfetcher.html; feed-id=xxxxxxx)"
Cause
The problem was one user had configured a Chrome plugin for RSS with the url of his/her filter (/issues/?filter=nnnnn). This plugin on the browser provoked requests from FeedFetcher-Google.
Resolution
- Block the requests on Proxy or Web server.