Scraping multiple exporters via pushprox

[anthonyra]

I've looked around to see if there was documentation on this but couldn't find anything concrete with regards to pushprox. One of the suggested approaches would be to use relabel_config which drops everything that doesn't match source_labels... however I'm not sure if the labels is being returned when using the http_sd_configs approach.

Each metric collector will most likely need to be a separate job (to ensure that all targets have the specific metrics for collection) but would it make sense to use relabel_config or maybe implement a filter query parameter on the http://localhost:8080/clients endpoint.

[peekjef72]

Maybe it is too late for you, but...

The only thing you have to do is to specify the host:port of the exporter to scrap it:

• so client:9100 for node_exporter,
• client:9182 for window_exporter
• and so on...

relabeling is only required if you have "pollers" like for blackbox that act as central point for scraping.

scrape_configs:
- job_name: node
  proxy_url: http://proxy:8080/
  static_configs:
    - targets: ['client:9100']  # Presuming the FQDN of the client is "client" and exporter is listening on port 9101.
    - targets: ['win_client:9182'].
- job_name: job2
  proxy_url: http://proxy:8080/
  static_configs:
    - targets: ['client:port']  # Presuming the FQDN of the client is "client".

this acts like in cmd shell:

curl -v --proxy http://proxy:8080 \
    -H "X-Prometheus-Scrape-Timeout-Seconds: 5" \
    http://client:9100/metrics

curl -v --proxy http://proxy:8080 \
    -H "X-Prometheus-Scrape-Timeout-Seconds: 5" \
    http://win_client:9182/metrics