About this talk
This talk emphasizes the importance of observability and performance insight for applications running in Kubernetes environments. The speaker discusses how Cilium utilizes eBPF to gather observability data, including Prometheus metrics, without the need to modify existing applications. A demo application showcases performance metrics and their variation with increased request volumes, along with how new configurations can introduce error rates and impact request durations. The session also covers exporting tracing headers using Hubble HTTP metrics as Exemplars, linking metrics to traces in Grafana for comprehensive monitoring of HTTP requests and durations using Tempo. Attendees will gain insights into monitoring service-to-service connectivity and collecting tracing data and key metrics through Prometheus, Grafana, and OpenTelemetry with Cilium and eBPF.
