status-go/metrics
Igor Sirotin 679391999f
feat_: `LogOnPanic` linter (#5969)
* feat_: LogOnPanic linter

* fix_: add missing defer LogOnPanic

* chore_: make vendor

* fix_: tests, address pr comments

* fix_: address pr comments
2024-10-23 21:33:05 +01:00
..
node feat_: log on panic (#5849) 2024-09-27 06:37:32 +08:00
README.md add README for metrics (#1906) 2020-03-17 22:09:21 +01:00
metrics.go feat_: `LogOnPanic` linter (#5969) 2024-10-23 21:33:05 +01:00

README.md

Description

This package configures Prometheus metrics for the node.

Technical Details

We use a trick to combine our metrics with Geth ones.

The NewMetricsServer() function in metrics.go calls our own Handler() function which in turn calls two handlers:

  • promhttp.HandlerFor() - Our own custom metrics from this package.
  • gethprom.Handler(reg) - Geth metrics defined in metrics

By calling both we can extend existing metrics.

Metrics

We add a few extra metrics on top of the normal Geth ones in node/metrics.go:

  • p2p_peers_count - Current numbers of peers split by name.
  • p2p_peers_absolute - Absolute number of connected peers.
  • p2p_peers_max - Maximum number of peers that can connect.

The p2p_peers_count metrics includes 3 labels:

  • type - Set to StatusIM for mobile and Statusd for daemon.
  • version - Version of status-go, always with the v prefix.
  • platform - Host platform, like android-arm64 or darwin-arm64

The way this data is acquired is using node names, which look like this:

StatusIM/vrelease-0.30.1-beta.2/android-arm/go1.11.5
Statusd/v0.34.0-beta.3/linux-amd64/go1.13.1
Geth/v1.9.9-stable-5aa131ca/linux-amd64/go1.13.3

This 4 segment format is standard for Ethereum as you can see on https://ethstats.net/.

We parse the names using labelsFromNodeName() from node/metrics.go.

Links