Cloud-native environments consisting of microservices, containers and Kubernetes need a new way of doing application performance management. The environments are architecturally different from traditional environments and continually changing.
It is very important to have continuous automatic discovery and mapping with containers, because change is the only constant. The application performance management must automatically discover and map all application components and then update the map when changes occur.
Containers can spin up and decommission in an instant, so if the metric granularity is too long, it leads to missing data and, worse, unidentified problems. So you need a tool that does one-second metric granularity.
Notification has to be as near to real-time as possible, so that you can roll back or fix any problems immediately. Three-second problem notification means you do not have to wait to see if a new container behaves properly.
In these environments, full-stack tracing of every request is critical. Every request is important in its own way and every one has the potential to fail or degrade. Any request you do not capture could be a bad trace that you did not debug.
The scale, complexity and dynamic nature of containerised apps can be overwhelming. So you need to focus on what matters. Of all the data that is collected, filter out extraneous systems and requests so you can focus only on the systems you are investigating.
Thousands of containers each hold thousands of elements. No human could manage application performance on their own. The application performance management tool must automatically determine which key performance indicators (KPIs) to collect, when measurements indicate service incidents and the likely root cause of problems. This is where there is the need to leverage artificial intelligence so these tasks can be automated.
To allow your operations and development trams to operate effectively, look for tools that have all six of these features. Anything less will jeopardise your container projects. The product that can do this is IBM Instana.
To learn more about how Instana can help you achieve better service quality for your cloud-native applications and environments, contact Birgit Smythe at bsmythe@envisage.email or visit https://www.instana.com.
Share