See Inside
a JAM Network.

A practical course on the abutlabs observability stack: Prometheus, Grafana, Loki and Alloy, fed by each node's own metrics, its JIP-3 telemetry or its JIP-2 RPC. You start the stack, run a network yourself, and learn to read what it tells you: which node is behind, how long a block takes to be finalized, and why a run failed.

Start the course Jump to the case study

What you will be able to do

Run the stack. Start it with one command, check it, and know what each service does and which port it answers on.

Generate your own data. Start a network of JAM validators from public images and watch it report in: every lab works on the run you make, not on ours.

Get any client in. Through its /metrics, its JIP-3 telemetry or its JIP-2 RPC, labelled so one dashboard shows every client.

Read the dashboards. Know the question each one answers, including how long a block takes to be finalized, and write your own PromQL and LogQL when a panel is not enough.

Diagnose a failure. Follow a real failing soak from its symptom to its cause (the refine time per package peaked at 13.2 s; after the fix, 0.60 s), then break your own network on purpose and find it the same way.

Learning Tracks

The stack

The observability stack is open source: abutlabs/observability (MIT). The networks in the labs come from jamswap, whose lasair6 net runs six lasair validators from public images. Telemetry follows the JAM Implementer Proposals (JIPs): JIP-2 for the node RPC, JIP-3 for telemetry.