Skip to content

Why does the federate endpoint return all metrics as untyped? #5589

@SleepyBrett

Description

@SleepyBrett

Proposal

Use case. Why is this important?

I would like to export some metrics from a prometheus server upstream to datadog, I'm attempting to use the veneur-prometheus scraper to accomplish this.

Unfortunately it fails because all the metrics return as "untyped".

I found this related issue: #3935

But i find the response condescending and without explanation, the article linked does not seem to answer the question.

Why are all metrics on the federation endpoint typed 'untyped' I assume the types are stored in the prometheus database. I'm only looking to federate a very small number of metrics ( around a dozen.

Is the type not stored and therefore not passable?

“Nice to have” is not a good use case. :)

Bug Report

What did you do?

Call /federate with a narrow query that returns a handful of metrics, but it's description blocks declare the timeseries as 'untyped'

What did you expect to see?
I expect the timeseries types to show up on the federation endpoint.

What did you see instead? Under which circumstances?

'untyped', always.

Environment

Prometheus 2.2.1 on kubernetes

  • System information:

NA

  • Prometheus version:

v2.2.1

  • Alertmanager version:

NA

  • Prometheus configuration file:
NA
  • Alertmanager configuration file:
NA
  • Logs:
insert Prometheus and Alertmanager logs relevant to the issue here

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions