Just because you have collected a few gigabytes of monitoring data, do you know anything about the health of your application? Monitoring is about more than just collecting data. Data is not information, it's just data. It is only useful once it's processed to turn it into information. A simple example would be, collecting time series data for "User CPU %". It's just data. Only once a rule is applied "If User CPU % > 80% then alert - User CPU load is too high, consider shutting down some processes" does the data transform into information. There is just way too much data available via monitoring of modern distributed systems, it's no longer consumable by humans, there must be some form of processing to generate actionable information from the data. What approaches may be used achieve this?
Steve is a Technical Marketing Manager at Instana, helping to transform Application Performance Management (APM) with AI powered automation. He brings plenty of experience to this role having worked in technical roles in the APM space since 2004 for companies including Wily Technology and AppDynamics. In that time has seen numerous technical revolutions and market disruptions giving him a unique insight into the rapidly changing DevOps environment. He has a background in electronics and programming before moving to software solutions.