Operational Analytics and the DevOps Journey

Are you deploying your application to production and finding out later that you need to rework your code due to an outage or errors?

This is a passionate topic at Data Mavens Ltd, which I have worked on for years, defending and finding ways to ensure that deployments can have a high degree of success. Where possible, I have embarked on huge efforts to train, document and lead my teams towards this excelling levels of success.

Error rates are not a pleasant experience, monitoring solutions often shine with the likes of BMC's Control-M showing how the automation of your workloads can bring faster response times to finding errors.

Successfully integrating Splunk into your monitoring and making the effort to use it allows your team to see the realtime effects of errors in your operational pipeline.

I personally remember a time when after a deployment Splunk was revealing errors that were so deeply embedded in the execution of workflows in an environment where the average executions in a day are about 10000; The range of workflows triggered per day were about 1500.

That was a good lesson for me to go back to the monitoring-testing-and-deployment drawing board and drive the journey to a DevOps culture using the Operational Intelligence gained from the failures.

Rome was not built in a day...

There are ways to avoid such disasters in production in the long run, if you are far behind on your Devops journey and live in the real world, like most  organisations, then yes, the success factor depends on many or a combination of manual/automated steps and a lot of disconnected/disjointed processes. Errors are high in their numbers and deployments regularly are nightmares - although some do enjoy the thrill... Yet the point is to remain cost effective and efficient and keeping rework to a low minimum.

If you have however adopted and are on your Devops journey then I do hope you are enjoying a stream of pre-prod testing that feeds into a seamless production deployment, using tools like Docker [Enterprise], Ansible, Jenkins, Rundeck, with your own home-grown playbooks.

Knowing in advance the impacts of your new build is the goal on every level. This is how you contribute to enjoying a higher success factor for your deployments.

"A journey of a thousand miles begins with a single step" - let that remind you that as long as you are taking these steps, you are going to get there. Just focus, assess and continue that curial path toward the change. Monitor these metrics and you will gain the drive that keeps you going in that right direction.