Building Resilient Continuous Integration Infrastructure with CloudBees - Webinar Q&A

Written by: Harpreet Singh
4 min read

I presented a webinar on Apr 24 about using CloudBees products to build a resilient CI infrastructure with Jenkins. I demoed building a reference architecture [1] with Jenkins Enterprise and Jenkins Operations by CloudBees products.

You can find more details about the products here [2].

Here are the questions from the webinar:

Q. Is the long-running builds plugin available today? Can you provide more details about the feature?
A. Jesse Glick, our lead engineer who built the feature, has blogged about it [3].
The long-running build plugin is available to existing Jenkins Enterprise by CloudBees customers through the experimental release update center.

Q. Why is there a degree of separation from the JOC server via the proxies? Or why are you using HAProxy here?
A. The idea is that users of Jenkins should only have to know one URL to go to. So, for example, Jenkins is set up as jebc.mycompany.com, primary instance is setup at jebc-host-a.mycompany.com:8080 and backup instance is setup at jebc-host-b.mycompany.com:9090. Without HAproxy, the end users will need to know the actual hostname and will have to switch back to hostname when primary goes down. With haproxy (or any reverse proxy), haproxy will route the request appropriately.

Q. What is the plugin used for backing up Jenkins?
A. A very intuitive plugin named "Backup plugin." :-) More details on the documentation page [2].

Q: Will Workflow be available to Jenkins Enterprise by CloudBees and to OSS users, as well?
A. The Workflow feature is a leap forward in Jenkins, one that will require us to make changes to the core. We will release core functionality in OSS and encourage plugin developers to extend it. For Jenkins Enterprise by CloudBees, we will look at building features on top of the core functionality. We did something similar with Folders, which is open source, while RBAC and Folders Plus play well with it and build on top of it.

As soon as the core underlying pieces are implemented and ready to be in a shareable state, we will release it in the community.Q. There are many plugins, what are the standard set of plugins and why is it not made standard with the Jenkins installation like RBAC, Backup, agent...It is hard for a new user to know what is required.
A. The plugins are bundled as part of Jenkins Enterprise by CloudBees. The plugins work with other scenarios beyond what I showed today (for example git-validated-merges plugin). Thus, you will need to install the ones that you care about.

Q. Can you comment on reducing downtime?
A. This was asked early in the presentation, and I hope that the webinar clarified it. The short answer is: setup individual teams to a controller, thus controllers end up being not heavily loaded (so, for example, no longer 20k jobs on one controller), thus increasing mean time between failures and reducing time to recover from a failure. The reference architecture document [1] talks about this as well.

Q.Thank you very much. This was great. I was wondering if you have another presentation like this for a company like ours that is using hosted Jenkins on Cloudbees already.
A. Sure, we have done a number of demos in the past that show various features like mobile development, continuous delivery with CloudBees. You can find the details here . If you will like to have a specific conversation about your use case, reach out to info@cloudbees and mention that you were on the webinar. Someone from our team will pull in an appropriate expert to have a conversation with you.Q: Can I use a Custom Update Center also to distribute tools installations across controllers/agents?
A. You can use it to distribute tools to controllers. More information on the documentation page [2].

Q: Why are most pages coming up as 504 and need refreshing?
A. Depends on the timeout of haproxy. Usually, WebEx sucks the juice out of my system, which had three Jenkins installations (plus builds), Skype running on it. Since my machine was low on memory, Jenkins was responding slower than the timeouts set up. I am going to do the next one on a machine on EC2 and leave my laptop with WebEx only.

Comments

Sounds great will do. Thanks, really great presentation. Also would like to say your service is the best, we have been using now for the last four months after switching from AWS.
Aww. thank you :-)

  • Harpreet Singh
    Senior director, product management
    @singh_harpreet

Links
[1] Reference Architecture for Jenkins
[2] Jenkins Enterprise by CloudBees (documentation , marketing page ) and Jenkins Operations Center by CloudBees (documentation, marketing page)
[3] Long Running builds blog (release notes )

Stay up to date

We'll never share your email address and you can opt out at any time, we promise.