-
Notifications
You must be signed in to change notification settings - Fork 3k
Closed
Labels
bugSomething isn't workingSomething isn't workingneeds triageNew item requiring triageNew item requiring triage
Description
Component(s)
exporter/influxdb, receiver/influxdb
What happened?
Description
The InfluxDB/otel conversion library assumed that OpenTelemetry histogram buckets are cumulative, like Prometheus histogram buckets. The result is that the bucket counts are always incorrect after conversion.
Originally reported here: influxdata/influxdb-observability#60
Steps to Reproduce
Either:
- Configure otelcol-contrib to export InfluxDB/Telegraf metrics with Prometheus v1 or v2 schema
- Receive an OpenTelemetry histogram via OTLP
Or:
- Configure otelcol-contrib to receive InfluxDB/Telegraf metrics with Prometheus v1 or v2 schema
- Receive a histogram from Telegraf via line protocol
Expected Result
This otel histogram:
0.05: 24054
0.1: 9390
0.2: 66948
0.5: 28997
1: 4599
inf: 10332
Yields this line protocol (schema is Telegraf Prometheus v1 in the example):
count=144320,sum=53423,0.05=24054,0.1=33444,0.2=100392,0.5=129389,1=133988
And visa-versa.
Actual Result
This otel histogram:
0.05: 24054
0.1: 9390
0.2: 66948
0.5: 28997
1: 4599
inf: 10332
Yields this line protocol (schema is Telegraf Prometheus v1 in the example):
count=144320,sum=53423,0.05=24054,0.1=9390,0.2=66948,0.5=28997,1=4599
And this line protocol:
count=144320,sum=53423,0.05=24054,0.1=33444,0.2=100392,0.5=129389,1=133988
Yields this otel histogram:
0.05: 24054
0.1: 33444
0.2: 100392
0.5: 129389
1: 133988
inf: 144320
Collector version
v0.73.0
Environment information
Environment
OS: MacOS 13.2 (and surely every OS)
Compiler(if manually compiled): go 1.19.5 (and surely every compiler)
OpenTelemetry Collector configuration
No response
Log output
No response
Additional context
No response
Metadata
Metadata
Assignees
Labels
bugSomething isn't workingSomething isn't workingneeds triageNew item requiring triageNew item requiring triage