waku-org/nwaku

chore: add dashboard for liteprotocoltester runs

Opened this issue · 1 comments

Background

As closing the light protcol tooling we shall add grafana dashboard for liteprotocoltester to be able to monitor the performance of the tests.

Details

As it is already possible to run liteprotocoltester on waku-simulatior or on any real fleet either on TWN.
Staticsitcs are also collected but only printed into the log.
We need to introduce corresponding metrics and prepare a dashboard for them.

Acceptance criteria

Run liteprotocoltester on TWN or on simulator, check monitoring dashboard and take screenshots of it.

  • run the tool on status prod fleet with accompanying dashboard
  • attempt to be close to a status light client behaviour:
    • use a couple filter and light push node
    • use a store node
    • do a small number of regular queries (not stress test but monitor)
    • restart node every hour or 30min (as would a mobile or desktop do).

Monitor metrics to get health of the network:

  • message reliability stats: confirm number of message sent through
  • connectivity: number of connected peers for the various protocols
  • peer exchange: report number of nodes returned from peer exchange and their relevant (shard, light push protocol support)

This way, we can have a better idea of the network health:

  • is a new client coming in able to make a libp2p connection to a light push, filter and store server?
  • if not, why ? (not relevant nodes in peer exchange response, not able to connect?)
  • is a new client able to send messages and receive them?
  • etc