We have developed an automated issue creation/update tool based on both an email gateway and a web service API. Issues are updated using the C# Atlassian SDK.
We noticed bad performance in the tool recently and found out that it is actually the call to "issue.SaveChanges();" that takes almost a minute to complete. That seems strange because this also holds if we only update a single field. And although that means we can only update one ticket per minute, the JIRA instance seems to be unusable also for other users at that time.
Support told us that a number of calls to
/rest/api/2/search
would cause an out of memory problem and that we should switch to OAuth.
Are there known problems using username / password authentication? And if so, does that explain why ticket creating takes so long?
--------------------------
Edit: Sorry, I have to be more clear about our problem.
Searching for issues by a custom field takes two seconds
2017-08-17 01:23:35,311 DEBUG API Find Issue by Hash
2017-08-17 01:23:37,518 DEBUG API Find Issue by Hash, Success = True
and updating the issue afterwards takes over 30 seconds
2017-08-17 01:23:37,518 DEBUG API SaveChanges
2017-08-17 01:24:11,861 DEBUG API SaveChanges returned
Atlassian support told us they don't help with development topics but that maybe switching to OAuth would help. I doubt it because I would then assume search to be equally slow...
It is slower, because it's doing all the hashing and user validation every time. Oauth enables the work to be done with internally cached credentials which does speed it up.
Thank you. I have updated the question because it was not clear that this is not a theoretical question but based on a actual performance problem that your answer does not explain for me.
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.