Use job and __param_<name> labels cause double instance in report

[stimg]

v. 0.17.0
Using config like this

[{
    "targets": ["192.168.0.2"],
    "labels": {
        "job": "my_job",
        "__param_module": "my_module",
    }
}]

in the targets file "my_job.json'" cause creation of two instances with the same parameters in the same job in the Prometheus metrics report. Workaround: delete job label.

[brian-brazil]

Can you share your prometheus.yml?

[stimg]

global:
   scrape_interval:     15s # By default, scrape targets every 15 seconds.
   evaluation_interval: 15s # By default, scrape targets every 15 seconds.

rule_files:
   - 'rules/*.rule'
   - 'rules/s3/*.rule'

scrape_configs:
...
  - job_name: 'dns_exporter_snmp'
    file_sd_configs:
    - names: ['/opt/prometheus/tgroups/snmp*_targets.json']
    relabel_configs:
      - source_labels: [__address__]
        regex: (.*?)(:80)?
        target_label: __param_address
        replacement: ${1}
      - source_labels: [__param_address]
        regex: (.*)
        target_label: instance
        replacement: ${1}
      - source_labels: []
        regex: .*
        target_label: __address__
        replacement: 10.200.18.51:9116  # SNMP exporter.

UPDATE: despite I delete job label, Prometheus creates two jobs for the target:

1. job_name: dns_exporter_snmp
2. job_name: my_job

[brian-brazil]

You're not doing anything odd, so that sounds like a bug that we'll need to verify and look in to.

I presume there's no other files matching that glob? That could explain it.

Does this also happen with 0.18.0?

[stimg]

uiuiui, you're right! there was another definition for this target hardcoded in the config. sorry.

[brian-brazil]

No problem.

[lock]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.