1

My goal is to tail the output and setup some health checks to monitor the nodes health.

So I'm looking for example outputs and actions to take. Should I search SRC for log outputs? Is there a best practice article for node health checks?

E.g.

  • Panics/non zero return = restart process if
  • uptime GT hrs else check hw spec
  • Target - finalised GT 10 blocks = check iops
  • Stuck at block X = bump version
  • Etc etc
2
  • For the "stuck at block", I'm assuming keeping updated to the latest will be fine. All host functions are there and sync is done via WASM. This should only be an issue if the node does not follow the releases.
    – Luna
    Apr 5, 2022 at 5:30
  • I think the node already supports some telemetry, did you look into that? wiki.polkadot.network/docs/… There is probably also a docker quickstart. Apr 5, 2022 at 12:20

1 Answer 1

3

I would start with this guide https://wiki.polkadot.network/docs/maintain-guides-how-to-monitor-your-node

Substrate nodes have builtin Prometheus exporter. Blockchain node usually runs without any interruptions for years or so, unless there's a need to update binary.

  • For collator nodes I would pay attention for block time and block authorship
  • For RPC nodes Its better to monitor WS connections, WS flags, consistency of file system. Most of these metrics are already implemented into Prometheus node exporter https://github.com/prometheus/node_exporter

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.