It does not take much to understand the benefits of the DevOps culture, processes, and tools. However, implementing DevOps in your organization is not as obvious and usually involves more than simply setting up tools. You have to convince team members, map old processes to new, and maybe even change the structure of organizational reporting and budgeting.
Unfortunately, there is no magic formula for implementing DevOps in an organization, but there are some strategies to help.
One proven strategy to adopt DevOps is to leverage log analysis.
Most technology implementation problems actually end up being people problems.
When you watch a series of failed projects, you start to see a common trend that usually involves the relationship between users, implementers, and decision makers.
Users often ask for functionality and don’t understand (or even care) how it is executed. Implementers are stuck holding the bag to get it all done, but not directly tied to the benefits. Decision makers are so detached that they can only look at the hard numbers like time-to-market and ROI.
The cultural side of DevOps tells us this is the wrong approach.
There should not be a silo between users, implementers, and decision makers. But, this is a surprisingly common problem, even in high-tech companies. Operations teams are implementing tools for development teams, which were bought by someone else.
The disconnect between these teams result in failed projects.
Fortunately, operations and dev teams share some common ground.
They both want to move fast and both are focused on results.
Operations does have the added burden of maintaining up-time, which can sometimes take their focus away from results.
Both also share a love for data.
Today it is hard to find someone who does not love digging into analytics and dashboards. The insights and visibility that modern analytics platforms provide can cast a spell on users.
Human nature compels us to learn more with less effort. This means there is a window of opportunity for organizations to implement and encourage adoption of DevOps with data.
Data in the modern software delivery pipeline can be used to know when something is wrong, when something can be improved, and what users like and don’t like. You could say that the DevOps practice is data obsessed. The mechanism for getting at this information is a robust log analysis platform; logging of the systems, applications, and even technical support and project management platforms.
Taking this same analytics platform, which will become the heart of your future DevOps practice and making it a tool right now for your data obsessed operations and dev team, will make the move to DevOps much smoother.
Here’s how you can make the move to DevOps easier:
Start early with a common pool of data.
You do not need to wait. Identify an existing pool of data that the organization desperately wants to make more sense of and read that data into a log analysis platform. You can then build a standard set of dashboards that answer common questions. As soon as you share this with the team, watch out; they will immediately double-click into the data and get addicted to seeing what else they can learn. You can do this without even implementing a continuous integration or delivery pipeline.
Invite data to your meetings.
The new log analysis platform should become a common character in all meetings regarding development and infrastructure. The dashboards should be exported and put in presentations with discussion around their insights. The organization will quickly realize that issues are identified faster, response time to issues is faster, and everyone can get on the same page of what is going on without a huge amount of effort.
Share and Share often.
If you configure automatic alerts or periodic reporting of beneficial reports and data via email to common users, you will notice a huge interest in the method of this reporting. You do not need to wait for a request unless there are security constraints; oversharing at the beginning is OK. Be careful thought as you might become so popular that the demands on the types of alerts and dashboards will become its own task. And later when processes are mature you will want to cull the alerts to be more deliberate.
Once everyone is addicted to the tool that provides them this valuable information, they will be hunting ways to get more data into it in order to get more insights. This naturally leads to a DevOps culture and eventually to a practice where teams move quickly and rely on data to tell them how they are doing. This culture also makes them less afraid to break something knowing that they can learn and move on from any failure quickly (“Fail Fast”).
Log analysis alone is not DevOps, but it is at the heart of any complete DevOps practice. Because log analysis is interesting to nearly all users in the organization it is the best place to build interest, culture, and practice for DevOps- even before it is implemented.