Hi,
I'm trying to make a simple JIRA search through api.
https://jira.mycompany.com/rest/api/2/search?jql=project%3DOK+AND+type%3DService+ORDER+BY+created+DESC&startAt=0&maxResults=10&fields=summary
When I execute the above GET on the test instance it takes 3-4 seconds to return data (only user logged on is me). However exactly the same GET on production instance takes ~25sec during peak traffic (5-10% server CPU used, 50-100 active users) and no less then 12-13 seconds during low traffic (after work hours, users mostly inactive but left logged on, <1% CPU used).
In terms of JIRA configuration both instances are identical - we have made a fresh restore of production instance on test environment.
Why there are so huge differences in REST API response time depending on traffic? Is there any way to lower the response time?
Community moderators have prevented the ability to post new answers.
Are you running it on Linux or Windows? A linux load of 5-10% is A LOT, while 5-10% on windows is barely nothing. Performance measuring and tuning Tomcat (and especially Confluence) is a lot of work but i took away a lot of good information on functions that tend to bottleneck the application-servers from this article (webserver is having issues now so link is to google cache)
Original link was: http://blogs.onresolve.com/2009/04/how-to-make-jira-6-12-times-faster/
In terms of JIRA configuration both instances are identical
If both servers are identical and you are comparing reponses between loads, of course you are going to see differences whether big or small. You'll need to figure out whether it is a network issue or whether it is pure load which is causing the slower response. For pure load, you can analyze the access log.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.