Description
Is your feature request related to a problem? Please describe.
We would like to be able to monitor success and failures of requests from Validator Client to Beacon Node.
Describe the solution you'd like
In the same way we can measure requests made by BN to ELs, we want to measur VC to BNs requests:
> curl -s 0:9200/metrics | grep engine_api_timeouts_total
engine_api_timeouts_total{url="http://erigon-01:8551",request="newPayload"} 1.0
engine_api_timeouts_total{url="http://geth-01:8551",request="newPayload"} 5.0
> curl -s 0:9200/metrics | grep engine_api_responses_total
engine_api_responses_total{url="http://erigon-01:8551",request="chainId",status="200"} 1.0
engine_api_responses_total{url="http://geth-01:8551",request="getBlockByHash",status="200"} 727.0
engine_api_responses_total{url="http://erigon-01:8551",request="newPayload",status="200"} 14284.0
engine_api_responses_total{url="http://geth-01:8551",request="chainId",status="200"} 1.0
engine_api_responses_total{url="http://erigon-01:8551",request="forkchoiceUpdated",status="200"} 14285.0
engine_api_responses_total{url="http://geth-01:8551",request="forkchoiceUpdated",status="200"} 14285.0
engine_api_responses_total{url="http://geth-01:8551",request="getBlockByNumber",status="200"} 40342.0
engine_api_responses_total{url="http://geth-01:8551",request="newPayload",status="200"} 14284.0
engine_api_responses_total{url="http://geth-01:8551",request="getLogs",status="200"} 11839.0
Except we do not care about type of request as much as we care about response status, timeouts, and latency.
Describe alternatives you've considered
We could measure request failures on BN side, but that would not show connectivity issues like timeouts.
Additional context
We want to be able to alert on VC failing to call BN REST API. This includes request errors as well as timeouts.
The request field is unnecessary for us, we mostly care about url, and also status.