I’m exporting Fly.io metrics to SigNoz using the OpenTelemetry Collector, but I’ve run into an issue — all the metrics are being collected as gauges instead of counters. Because of this, I’m unable to use the SigNoz query builder effectively, as it doesn’t allow me to perform the expected operations on these metrics.
Could you show a sample of the data you’re sending? AFAIK, time-series data doesn’t tend to be sent as a gauge or a counter; that’s generally a rendering decision on the remote side. Doesn’t SigNoz have graphing widgets, and you just need to choose the right widget for the job?
Yes, I am sending basic metrics for a simple app.
This is the otel.yaml I am using to scrape Fly’s federation endpoint -
receivers:
prometheus:
config:
scrape_configs:
- job_name: fly-federate
scheme: https
metrics_path: /prometheus//federate # e.g., /prometheus/personal/federate
params:
match: [‘{name=~“fly_.*”}’] # only Fly metrics
static_configs:
- targets: [“``api.fly.io``”]
authorization:
type: FlyV1
credentials_file: /etc/otel/secret/fly_federate_token
I then generate some traffic -
for i in {1..30}; do curl -sS https://<your-app-name>.fly.dev >/dev/null; done
And then metrics show up :
curl -sS -i -G
-H “Authorization: FlyV1 $(cat fly_federate_token)”
–data-urlencode ‘match={name=~“fly_.*”}’
https://api.fly.io/prometheus//federate`` | head -n 40
The only problem is it being exported as gauge ( even the ones having _count in the name ) -
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.
