In DotCi builds are still stored as xml in mongodb with some fields like 
pusher, sha, branch, status etc stored as json attributes for querying . 
I tried serializing builds into pure json instead of xml but ran into tons 
of issues with plugins and I was not sure what plugins were doing out in 
the wild so I stuck with xml.

We run ~3000 builds per day and we can't do aggressive pruning of builds 
because of SOX requirements, this caused us huge performance issues like 
startup time of over 50 minutes ( this was a while ago, not sure if lazy 
loading helps here). One option was that each team gets its own jenkins 
server but this was maintenance nightmare. 

Using mongodb and not loading all builds in memory all the time brought 
down startup time to within couple of minutes. Out Jenkins instance is 
fairly snappy despite having  100's of thousands of builds, for example 
 'my pull requests view' is a query to mongodb  on pusher field and doesn't 
involve traversing all the builds in memory. 


On Tuesday, May 20, 2014 11:36:30 AM UTC-5, Surya Gaddipati wrote:
>
> https://github.com/groupon/DotCi 
> github usename: suryagaddipati
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"Jenkins Developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to