Even for the most seasoned engineer, going to production can still bring about butterflies in your stomach. Before, during, and after your deployment a flurry of checks and activities to validate availability. Cracking open and tail’ing/grep’ing logs are the norm and during a deployment watching out for potential errors and leaning on your trusted monitoring solutions looking out for regression.
During a deployment, too often the availability of the application/change equals success. To get a true sense of the health and performance of a deployment requires potentially multiple engineers with domain knowledge of the infrastructure and application. On the other hand, your customers expect no regression or poor performance during deployments which warrants multiple engineers validating deployments.
What if the tribal knowledge held by the multiple engineers could be automated by AI/ML? This session will take a look at some of the machine learning techniques you can apply to automate deployment verification and health checks, specifically for enterprises utilizing Datadog.