Spun off from German IT services provider CodeCentric, Instana is one of the challengers of APM incumbents like AppDynamics and Dynatrace.

Instana uses an agent-based approach to collect infrastructure data as well as application flows, and make the collected data available in a SaaS only analytics and dashboarding system. In addition to collecting metrics on CPU, memory and other infrastructure data, the agent attaches itself to running application processes. For the applications it has instrumented, the agent will intercept incoming and outgoing connections (such as HTTP calls in and out, database calls, cache requests, ...) and use distributed tracing techniques to visualize all the collected data together in the form of application flows.

Instana uses a number of custom-built algorithms to detect possible issues and incidents, for example:

  • Sudden drops in request rates in services
  • Sudden rises in error rate
  • Disk volumes that will fill up soon given current usage rate

At time of writing (2/2019), Instana does not support customization of the incident detection algorithms, and is missing features for filtering and prioritizing the stream of detected issues and incidents.

Instana does not collect data from other monitoring or log aggregation systems. However, through its agent it also collects log statements from instrumented applications and collects a comprehensive set of metrics itself. It offers an API for querying the metrics it collected.