Thanks to the kindness from Gliffy, we can now add diagrams to Wiki pages, in a way that enables collaborative edits.
See more info, including a sample diagram in a Wiki page.
Over drinks this evening Kohsuke pointed out that he never saw a blog post about PagerDuty.
If you've never worked in a sysadmin role or in any other position that would require an on-call rotation, then you may have never seen PagerDuty.
In essence the service provides a great series of integration points with Pingdom or Nagios for handling monitoring. As an infrastructure guy (part time), I can honestly say it's a great tool and I'm grateful to PagerDuty for supporting Jenkins with our own account to help manage project infrastructure.
A couple weekends ago I finished setting up Nagios (read-only username/password: jenkins/jenkins) for critical project services which by itself is a good step forward. Combine that with PagerDuty's Nagios integration and a solid on-call rotation, and I'm more confident than I've ever been that Kohsuke or myself could actually take a vacation!
Check them out, and be sure to thank them on Twitter at @PagerDuty for supporting Jenkins!
Let me preface this entire post with this: I love Contegix.
While working on some infrastructure tasks I had long-since put-off for the Jenkins project, I noticed something this weekend that scared the hell out of me.
At some undetermined time, our MirrorBrain
installation stopped redirecting to our mirror network. Absolutely zero
downloads were being redirected, meaning that cucumber, the 1U machine
graciously colocated by Contegix had served up
far more bits than I ever wanted it to.
As such, I would like to publicly apologize to Contegix on behalf of the Jenkins project. Their support for the project has been tremendous but this glitch caused such an incredible amount of traffic to be pushed through their network that I feel exceptionally bad about it (turns out, Jenkins is pretty popular!)
Now, for the good news. In diagnosing and debugging this issue (in a caffeine-fueled frenzy I might add) I managed to do a couple things:
Within 30 minutes of correcting the error, I was able to add two mirrors in Germany, re-enable one from Taiwan and add a new mirror in Japan!
This post is long over-due and I really apologize for that.
Some months ago we put out the call for "more slave machines!" through the mailing lists, sky-writers and twitter. We had a serious problem, for a continuous integration project, a large number of our plugins and dependencies weren't being built in a continuous and automated fashion!
We had some builds on a couple of flakey machines on home connections contributed by various individuals, until Rackspace stepped up in a big way and donated an infrastructure server for the project to use.
For months now, just about all plugins and core have been built and tested on spinach,
the always-on machine in the Rackspace Cloud. Dutifully chugging away building
core, plugin after plugin and occasionally getting flooded with work from Frederic
Camblor's plugin compatibility tester!
In hindsight, having a powerful infrastructure machine for nothing other than builds has helped us build great software faster; I can't imagine how difficult things might be otherwise.
I've personally had a lot of interaction with Rackspace engineers through the OpenStack project and have a number of friends who operate businesses on Rackspace/Rackspace Cloud hybrid infrastructures.
The folks at Rackspace are top notch and I can't thank them enough for contributing to the Jenkins project.
In a nutshell Jenkins CI is the leading open-source continuous integration server. Built with Java, it provides over 400 plugins to support building and testing virtually any project.
