It's time to give the network the same home that applications already have. Grafana and Tempo are where your team already goes when something breaks at 2 a.m., so let's make sure that network data is easy to view in your Grafana dashboards. We will explore exporting pod-to-pod traffic as OTLP trace spans, landing them in Tempo through the OpenTelemetry Collector, and building Grafana dashboards with TraceQL for cross-namespace dependencies over time, traffic spikes, RST and retransmit patterns, and flow spans sitting right next to your app traces.

We will leave with a clear map of when to reach for Hubble and when to reach for Grafana, a Collector-to-Tempo pipeline you can run today, and dashboard patterns built for real on-call work, all utilizing the OTLP open standards throughout. Any team running Kubernetes and Grafana can use this, whether you are debugging your first Cilium cluster or operating network visibility at scale.

Target audience is platform engineers, SREs, and observability engineers who run Kubernetes with Cilium and a Grafana-centric observability stack. Attendees should be comfortable with OpenTelemetry concepts and basic Kubernetes networking. No deep eBPF or Cilium contributor experience required.

This will be the first public delivery of this session. It has not been presented at a prior conference.

Technical requirements include a projector or large display, HDMI or standard conference AV. I will provide a macbook pro with the dashboards and some historical data to showcase (no wifi needed). The live demo uses a web browser for Grafana Explore and dashboards.

All code and configuration is publicly available on GitHub, and is available for audit and review. This is a community technical session focused on promoting open standards in the network observability stack of open telemetry.