Putting the Ops in DevOps
It is 3 am in the morning and your phone rings. You think to yourself that it can either be a call from a drunk friend, but then you realise all your friends send texts, or deep down, you hope it’s your father, phoning to finally say that he is proud of you (I won that Eisteddfod for you Dad!). We all know that most of the time, it’s more a case of the Command Centre at work, phoning to say ‘Sh*t just got real’.
Download the presentation here.
At Africa DevOps Day June 2017, Kevin Rolfe expanded on the point that we should never forget to place the Ops into DevOps and mused about the fundamentals of the DevOps movement and what it all actually means.
He questioned if it was about people, processes, tools or the notion that we need to analyse everything to death? The idea that DevOps creates customer value and innovation is a lie, he stated profoundly. In reality, it’s about the customer experience. The question that we really need to be asking is how does the customer experience their entire journey?
Kevin went on to relate his own DevOps journey and said that the problem they faced was that DevOps could not scale. Every time they tried, they lost the Ops pipeline in the journey.
The focus seemed to always be in tech, rather than on ops. So, Kevin spent his time on architecture and design thinking. Through design, he focused on availability, planned for capacity, built consensus on important decisions and above all, kept it simple. He found that forced organic teams cannot scale effectively. There were no real, tangible metrics that captured the real numbers (see a slide of the 4 quadrants). He also found that you had to keep practising incident responses, rather than just pushing code and that the best way to solve theoretical arguments was through execution.
Test if it’s going to work first. In this way, the teams can see if they are having real world impact on anything and if they are following real Agile practices. Through these practices and metrics, it was easier to identify High Growth Teams.
In order to coordinate all these Ops, toolsets are not important, the ability to get the right focus in the room is more important. This can be achieved through bots, using intelligent plugins into APIs. This helps to make smarter decisions on your behalf. Use tools with robotic enablement and Machine learning that frees you up for exceptions and apply your expertise.
The key is to integrate your platform and your bot (make a robotic user with right permissions) with your Active Directory. In the end, you will have to connect your bot to a Bot Framework.
In the end, the critical thing to remember, according to Kevin Rolfe, is Dev. Ops. This would create customer value and innovation.
Kevin Rolfe’s Top Tips:
- Never forget the Ops. In DevOps.
- Stop focusing on Agile, Culture and Tools
- Put applications and infrastructure through the same workflow
- Plan for capacity
- Use scalable design
- Build consensus on important decisions
- Simplicity
Was this blog post helpful? Don’t you think your buddies would like to know more about what was learnt at Africa DevOps Day 2017? Please share on Twitter, Facebook & LinkedIn.