Albert Rigo shares his thoughts on the Atlassian Accelerator Day in Copenhagen, and how he learned to quickly and comfortably scale Atlassian Data Center, upgrade with zero downtime and stop spending his weekends upgrading JIRA.
Atlassian Accelerator Day in Copenhagen
A late bus and a cancelled train didn’t stop me from happily attending the Atlassian Accelerator Day in Copenhagen. I had been so excited to learn more about Atlassian Software in Kubernetes (ASK) that I had left home far too early.
The day had both a talk and demo on the agenda, and I would discover they were definitely worth crossing the Øresund for. The talk would be inspiring and get me excited for the demo, which would not disappoint.
Scaling your business with Atlassian Data Center
Henrik Høegh started his talk by pointing out the importance of being able to scale easily, and how Atlassian Data Center (ADC) can help you do so.
Nobody likes getting stuck in queues, and to scale with customer demand, supermarkets open or close extra cash registers. Atlassian Data Center tools follow the same principle. To meet load demands and scale, you spin up a new node, connect it to the load balancer and have it join the cluster without downtime.
While users might not notice the difference, as everything just working quickly becomes the norm, introducing ADC will definitely be noticeable for its maintainers. For starters, the pains of upgrading melt away. Atlassian Data Center offers zero downtime upgrades, a feature many “high-availability” tools fail to deliver.
Zero downtime upgrades are conceptually simple. You disconnect, upgrade and reconnect your nodes individually. While the nodes you’re upgrading are down, those that are still connected to the load balancer continue to serve the users.
He went on to describe many of the other great features ADC has to offer, such as its performance at scale and disaster recovery options. He also briefly mentioned the currently supported applications including JIRA Software, JIRA Service Desk, Confluence, Bitbucket and HipChat.
High availability and resilience
Henrik then explained how they added high resilience, configuration as code and many other benefits to their ADC setup. By combining Kubernetes and with the Atlassian Data Center, they had created a truly powerful and resilient setup: Atlassian Software in Kubernetes (ASK).
After the talk, Timothy Harris gave us a demo where he added new nodes to a running ADC cluster. With his Kubernetes-enhanced setup, he was able to spin up and connect new nodes with a single command. As someone who loves simple and pragmatic solutions, I was happy to find the final result to be so usable.
Here’s a quick overview of some of the tools involved:
Kubernetes adds resilience and self-healing, allowing nodes that die to be fired up again instantly.
Dashboard is great for giving you a visual overview of your Kubernetes cluster.
Træfik as the preferred load balancer, as it’s fast, lightweight and conveniently configured through the Rest API.
After the great talk, demo and some socializing, I headed home, tired but excited. Luckily, public transport was kinder to me on the way back. More than anything, I was happy to find the final result so simple. It reflects in my favorite quote of the day, by Timothy Harris:
“It’s like, one command, and it scales, which is really cool.”