Closed eranra closed 2 years ago
Could it be that the Collector Ingest is not really forwarding the metrics as long as they come regularly?
The way I interpret this line: https://github.com/netobserv/flowlogs-pipeline/blob/main/pkg/pipeline/ingest/ingest_collector.go#L153 is that the "flush timer" is restarted each time a new metric arrive, so it won't forward them unless there is 1 complete second without receiving metrics.
(EDIT: this will be fixed in an upcoming PR)
The bug seems to be in these lines: https://github.com/netobserv/flowlogs-pipeline/blob/e0d95f28d1f2973e52b260a73b1e750e5caec990/pkg/pipeline/transform/transform_network.go#L77-L80
The first %s
corresponds to outputEntries[rule.Input]
which is float64 (not a string). The error is
`Invalid token: '%!'`
But it's swallowed. @eran Is there a reason for not logging the error?
A simple solution would be to change it to %v
. I'll submit a PR for that...
Initial investigation shows that we do not create the metrics