You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
As part of investigation by @mattmathis to find a method for instrumenting ndt with application metrics we found that some ping/pong rtt metrics are being generated but not collected during ndt-server measurements.
To determine if this functionality is used by all clients and to get a sense of the distribution of these values, we should add prometheus histogram that collects the application rtt metrics. Labels should include: direction, protocol (and exclude monitoring measurements).
Based on the counts, and values, we can get a hint about how many clients support this and whether it may be worth archiving this metric.
The text was updated successfully, but these errors were encountered:
BTW the echo function is generally part of the base spec every IETF inspired protocol. I doubt you will find any implementations that don't support it.
As part of investigation by @mattmathis to find a method for instrumenting ndt with application metrics we found that some ping/pong rtt metrics are being generated but not collected during ndt-server measurements.
The rtt values are logged but not archived.
https://github.com/m-lab/ndt-server/blob/master/ndt7/receiver/receiver.go#L42
To determine if this functionality is used by all clients and to get a sense of the distribution of these values, we should add prometheus histogram that collects the application rtt metrics. Labels should include: direction, protocol (and exclude monitoring measurements).
Based on the counts, and values, we can get a hint about how many clients support this and whether it may be worth archiving this metric.
The text was updated successfully, but these errors were encountered: