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
We have seen cases where we need to collect telemetry statistics during the course of a track execution. For example, we might want to use previously collected metrics to compute a statistic - that can't be easily calculated in Kibana. For now, we do this work outside of the track.
Currently neither runners or param sources give us access to the metrics store. The proposal is to provide an interface, similar to that available for runners and param sources, where we can query and index into the metrics store - presumably this would be restricted to when es is enabled as the metrics store.
When writing a new telemetry device, we need three steps:
write to the metrics store (done in the actual telemetry device class)
transform the data to aggregate it to appear in results as a single data point (done in GlobalStatsCalculator)
print the results in the command line report by reading data from step 2 and format/compare them in a human-friendly way (done in SummaryReporter and ComparisonReporter)
This issue is only about 1. which is why @danielmitterdorfer opened #1314 for 2. and 3., but we see that we can't really dissociate the two issues, so with his blessing I'll close #1314 in favor of this issue.
We have seen cases where we need to collect telemetry statistics during the course of a track execution. For example, we might want to use previously collected metrics to compute a statistic - that can't be easily calculated in Kibana. For now, we do this work outside of the track.
Currently neither runners or param sources give us access to the metrics store. The proposal is to provide an interface, similar to that available for runners and param sources, where we can query and index into the metrics store - presumably this would be restricted to when es is enabled as the metrics store.
@danielmitterdorfer
The text was updated successfully, but these errors were encountered: