Observability

Platon's observability platform is built on the Grafana stack: Loki for logs, Mimir for metrics, and Tempo for traces. All three are accessed through grafana.platon.sikt.no.

There are two ways to use the platform: logging only, or full observability with OpenTelemetry. Both are fully supported.

Logging only

For teams that have a working logs-centric setup, or who are not (yet) adopting OpenTelemetry (OTel).

Start here if:

• You run on Platon PaaS and just want logging. See PaaS Logging.
• You run on Puppet-managed VMs. See Puppet VMs.
• You run outside Platon and want to ship logs in. See External Services.

Reference material: Logs, LogQL Best Practices, Alerting.

Full observability with OpenTelemetry

For teams that want spans, metrics, and traces in addition to logs. OpenTelemetry is the emerging industry standard for producing this telemetry, and Platon's observability stack receives it natively.

Start here if:

• You are new to OpenTelemetry and want a short orientation. See OpenTelemetry tutorial.
• You already know OTel and just need the Platon-specific setup. See PaaS Full Observability or External Services.

Reference material: Choosing your signals, Logs, Metrics, Traces, Alerting.

Kubernetes metrics on PaaS

Every Platon PaaS workload automatically emits Kubernetes metrics (CPU, memory, pod restarts) to Mimir, with no setup required. They appear in your team's Grafana organization and in shared dashboards in the "Sikt" Grafana organization. See What you get for free on PaaS.

Access Grafana

Log in at grafana.platon.sikt.no with your Feide credentials. You will land in your team's Grafana organization. If you belong to multiple, you can switch between them.