Observability
Because all mesh traffic flows through the data plane proxies, Istio can generate consistent telemetry for every service without each team instrumenting their own code.
What it Gives You
- Metrics - request volume, success rate, and latency for traffic between services, in a consistent format.
- Distributed tracing - trace context propagation so you can follow a request across multiple services.
- Access logs - per-request logging at the proxy level.
- Traffic topology - a live view of which services are talking to which, and how healthy those connections are.
How it Works
The proxies emit telemetry automatically. This telemetry is typically collected by Prometheus and visualised with tools such as Grafana and Kiali.
| Tool | Role |
|---|---|
| Prometheus | Collects and stores mesh metrics |
| Grafana | Dashboards for mesh and workload metrics |
| Kiali | Service mesh topology, traffic health, and configuration validation |
On AKS, the Istio add-on is verified to work with Azure Monitor managed service for Prometheus and Azure Managed Grafana .
Info
Tracing still requires applications to forward trace headers between inbound and outbound requests. Istio generates and propagates spans, but it cannot stitch a trace together if a service drops the trace context.
When to Use It
- Understanding service-to-service dependencies in a complex application
- Spotting latency or error-rate regressions after a deployment
- Troubleshooting where a request is failing across multiple hops
You will visualise mesh traffic with Kiali in Lab 4 .