Pledge

I pledge to pay $100 if the following gets implemented in its entirety:

Is your feature request related to a problem? Please describe.

I'm tweaking parameters, updating the relay, observing genuine demand fluctuation and probably DOS attacks and only have CPU load, Disk I/O, Bandwidth and some other parameters that are logged for the whole machine to judge if anything is wrong. That's not good for taking informed decisions.

Describe the solution you'd like

• Log key performance indicators
• Expose them through a nice interface (serverUrl/stats for example) maybe using grafana or similar

KPIs I would love to see:

• concurrent websockets open
• concurrent queries watched
• websockets opened/closed
• time from connect to first EOSE
• time from [e:[] query to EOSE
• events served
• Standard system load parameters: CPU, Load, Memory, Disk I/O, Disk Usage, Bandwidth

For some of these parameters, aggregate functions like median, 95th percentile ... would be of interest.