The moment you’ve all been waiting for…

[#1]

It’s finally here! The Data Center Deployments team is excited to announce that the Data Center Helm charts are now generally available.

You can start here.

If you aren’t familiar with Helm charts or the value that they provide, you can check out the first post in our series.

What does this mean for you?

Simply put - you can deploy Jira Software, Jira Service Management, Confluence, and Bitbucket Data Center on Kubernetes (K8s) clusters running on either your own hardware or a cloud provider through our customizable templates, and be officially supported by Atlassian. If you run into any snags or have questions, our support team is ready to make sure that your deployment is successful.

Of course, we’ve also created detailed documentation that will help walk you through the process and answer any of your questions. We will continue to update it to ensure that you stay supported. If you are interested in technical details, feel free to explore the source repository.

Big thanks to all of you

We wanted to take a moment and just say a big thanks to our Community members who tried the Helm charts as part of our beta program. Your feedback was incredibly valuable to our team.

We also wanted to thank the pioneers who were already using Kubernetes to run their Data Center products. Your insights and considerations have helped us as we’ve worked on officially supporting K8s.

And most importantly - we wanted to thank all of our members. The goal of this space has not only been to get y’all excited about the Helm charts but to create a space where we could come together and learn from each other.

We’ve got more posts and videos coming out in the next few weeks, so stay tuned!

[Jan B.]

Great work so far. It's so nice to have the helm charts and the documentation is really helpful.

What I would really like to see is a working liveness probe for the Pods. It doesn't happen really often, but I already experienced Jira and Confluence crashing internally. In those cases it would be helpful if Kubernetes is able to recognize it to kill that pod and create a new one.

Using the /status-Page there, too, leads to the Pod being restarted when doing a full reindex since it will get a 503 Response during that. By now I wasn't able to find a solution that works without any drawbacks.

[Dylan Rathbone]

Hi @Jan B_ ,

Thanks for the feedback! With regards the liveness probe, that is something we had considered, and may well be something we come back to in the coming releases.

As you've pointed out when performing a full-reindex, a liveness probe will restart the pod on account of the 503. Worse still, this may cause an indefinite loop of pod restarts. We're currently exploring options for the status endpoint itself, especially with regards how it reports the current state of the index. Based on this we may be able to introduce a liveness probe at a later date.

In the meantime I'd be keen to understand why Confluence and Jira is intermittently crashing for you? Any additional details would be great.

[Bryan Guffey]

Hey y'all - Can we use these for our Server products as well? Thanks!

[Dylan Rathbone]

Hi @Bryan Guffey, thanks for the question.

In short no. To realise the benefits that K8s provides through a clustered environment the Helm Chart's utilize the existing Atlassian Data Center Docker images.

Thanks,

Dylan.

[rammohan.mididoddi]

Curious to know which latest version of Jira supported these docker images?

[Idan Bidani]

Does the docker images require privileged permissions to run?

[Dylan Rathbone]

@rammohan.mididoddi if talking about the Helm charts, the minimum versions that we support for each product are described here i.e

Jira: 8.19

Confluence: 7.13

Bitbucket: 7.11

[rammohan.mididoddi]

@Dylan Rathbone 

Thanks for your info.

[Laura Holton _ACE Rotterdam_]

Awesome, thanks @Gaby Cardona 🙌

[Pramodh M]

Yay!! Thanks @Gaby Cardona 

Now the work is convince this is a good way to deploy the Application in the cloud if the customer is going with the Data Center model.

[Jorgen Christiansen]

Do you have any updates on the Jira Scaling issues being resolved around the Index process ?

[Dylan Rathbone]

Hi @Jorgen Christiansen ,

If you're referring to the indexing issues highlighted as part of the Jira limitations, then at this stage unfortunately not. Indexing issues such as these are something that the Jira team are well aware of and will endeavour to address based on their current backlog and priorities. We do have a documented workaround for these problems which I understand is not ideal, but please let us know if this isn't quite hitting the mark for you.

Thanks.

[Andrzej Kotas]

Hi Jorgen, please let us know which index issues you had in mind. If not the ones listed, could you share some more data around your experience? Always looking for more insights as we sort out the current backlog and priorities.
Thanks,
Andrzej

[Karen Mixon]

Hi there,

Any update on the liveness probe and the possible infinite loop of pod restarts? We ran into then when standing up Jira using the Helm chart.

Thanks!

Karen