Containers vary in memory size depending on the applications and microservices they support – and that raises a challenge. Deep instrumentation can be too expensive, while simplistic mapping can’t deliver the detailed performance insights needed for complex distributed systems integrated with existing enterprise applications. And API-centric communication also means needing to quickly determine how latency impacts application performance and end-user experience.
The key is having more than one way to monitor containers – without having to increase tools and cost. This means adopting non-intrusive agentless technologies that automatically discover, monitor, and trace all types and sizes of containerized applications and transactions, providing insights into latency speeds, error rates and end-to-end application performance. And the more auto-discovery capabilities the monitoring system has, the less that operational support is required.