As mentioned in The Beginning, the Atlassian setup that was inherited started out as 1 team's answer to a problem. As such, the infrastructure too was built for 1 teams answer. This included servers for Jira, Confluence, Crowd, and Bamboo / Fisheye / Crucible on a shared server. By the time my team became owners, Bamboo was no longer in use, Fisheye/Crucible was managing 15 XXX repos, Confluence had 20 spaces and Jira was sitting on ~100 projects (each with their own special snowflake configuration).
This first thing to address was the infrastructure.
After meeting with the Atlassian SME we brought on staff. It was recommended that Bamboo and Crowd be shut down (we were integrated with AD directly) and the other applications would all be put on the same server.
Let me pause here for a minute. If at all possible AVOID DOING THIS! (i.e. putting all the apps on one server). I was new and naive and went along with the recommendation. That decision alone cost us a few months of this journey and…well….I’m preaching to the choir now. You probably know better, but if you don’t know, DON’T DO IT! Maybe I'll write a whole separate blog on why this is bad, but for now let's get back on track.
So there we were, unused applications shut down, all in use apps on the same server and a clearer picture of what we were up against. This migration gave us our first chance to put our stamp on the Atlassian setup. We developed the communication strategy for server maintenance along with the cadence, we got noise (i.e. Crowd and Bamboo) out of the way, and more importantly we now had the applications on servers that fell within our ownership with administrative control to only our team.
This also put the rest of the rest of the organization on notice: A new team was now in charge and a focused effort would now be dedicated to the tool.
This also allowed for an easy segue to the next stage of our transformation, The Culling.