Observability

Observability is how you understand what Fission and your functions are doing in production — through metrics, traces, and logs.

Fission emits all three signals from its core components and from your function pods, in formats that standard cloud-native tooling already understands. This section shows you how to wire each signal into a backend and visualize it in Grafana.

The observability stack

Fission’s components and your function pods are the sources. Each source emits one or more signals, which a collector ingests and stores, and which Grafana then queries for dashboards and ad-hoc exploration.

flowchart TB
  subgraph sources["Fission Sources"]
    components["Core components<br/>(router, executor, ...)"]
    fnPods["Function pods"]
  end

  subgraph collect["Collectors / Stores"]
    prometheus["Prometheus"]
    otel["OTel Collector"]
    loki["Loki"]
  end

  jaeger["Jaeger"]
  grafana["Grafana"]

  components -->|"metrics"| prometheus
  fnPods -->|"metrics"| prometheus
  components -->|"traces"| otel
  fnPods -->|"traces"| otel
  components -->|"logs"| loki
  fnPods -->|"logs"| loki

  otel -->|"exports traces"| jaeger
  prometheus -->|"datasource"| grafana
  loki -->|"datasource"| grafana
  jaeger -->|"datasource"| grafana

  class components fission
  class fnPods pod
  class prometheus,otel,loki,jaeger store
  class grafana user
  classDef user fill:#ffffff,stroke:#94a3b8,color:#1f2a43
  classDef fission fill:#e8f0fe,stroke:#2d70de,color:#1f2a43
  classDef pod fill:#e6f7f1,stroke:#11a37f,color:#1f2a43,stroke-dasharray:5 3
  classDef store fill:#fff7e0,stroke:#dba514,color:#1f2a43,stroke-dasharray:5 3

Signals and guides

• Metrics — Fission exposes Prometheus-format metrics from every component and function. See Metrics with Prometheus.
• Traces — Fission instruments request flows with OpenTelemetry and can export to any OTLP backend such as Jaeger. See Tracing with OpenTelemetry.
• Logs — function and component logs can be aggregated with Loki and queried in Grafana. See Logs with Loki.
• Service mesh — Linkerd can mesh function and Fission pods to add request-level metrics. See Observability with Linkerd.