Team of Teams

What initially took 5 business days and 10+ developers to deploy to production was eventually done automatically in a few hours, 2 hours ahead of schedule the first time. When I was hired to join the team, every single senior developer with almost a decade or more of experience said it couldn’t be done. Since I had done it before, I chipped away at it and eventually with support from my manager, willing developers and the architecture team we got there. The only person who could understand how I felt was Andy Dufresne at the end of The Shawshank Redemption. After reading Team of Teams, I would call it my “awesome machine” but I soon came to realise that it wasn’t as awesome.

To give you an idea of the scale of the deployment pipeline, there were at least these environments

  1. 3 stand alone development environments
  2. 3 integrated ones
  3. 3 QA ones
  4. 2 performance testing ones
  5. 4 Virtual test environments
  6. 2 customer facing self-test environments
  7. 3 prod sized environments

How many VMs were there in a prod-sized environment? I remember the list having 200+ entries, this is what I remember base don my support responsibilities.

  1. The front end had 15 VM
  2. The API gateway had 4 VM
  3. The application tier had 20 integration servers, 10 MQs. Each IS had a MemCached server, there were 12 or more app servers
  4. There were 6 sets of 3 nodes of Oracle databases, so 18 of those
  5. 12 Oracle IDM related servers
  6. Then there were the Splunk servers and their predecessors.