Getting error in prometheus target snmp-exporter endpoint

[kalid-kto]

I have been doing tests to monitor a UPS but I get an error when trying to connect snmp mib scraping from the UPS trough snmp-exporter to prometheus...

In the error message shown by prometheus the URL is swapped to the one of the smnp-exporter

This is the url shown in connection refused error message in prometheus:
-- http://localhost:9116/snmp?module=MODULE&target=IP
And this is the acctual url from snmp-exporter:
-- http://localhost:9116/snmp?target=IP&module=MODULE

I tried several relabeling settings in the prometheus.yml configuration file, but I couldn't solve this problem.

Any idea what the problem is?

Thanks in advance for your suggestions.

[SuperQ]

Order in URL parameters doesn't have any effect.

[kalid-kto]

The problem was... that i didn't define correctly the job_name in prometheus.yml.
Prometheus was unable to connect to the SNMP exporter on localhost port 9116. This is likely because Docker Compose is assigning the SNMP exporter a host name. I should have used the name of the service, defined in the docker-compose.yml

I only change replacement: snmp-exporter:9116 instead of replacement: localhost:9116 and worked as I spected

- job_name: snmp-exporter
  static_configs:
    - targets:
      - ip.ip.ip.ip  # SNMP device.
  metrics_path: /snmp
  params:
    module: [ups]    
  relabel_configs:
    - source_labels: [__address__]
      target_label: __param_target
    - source_labels: [__param_target]
      target_label: instance
    - target_label: __address__
      replacement: snmp-exporter:9116 #Container_name defined in docker-compose.yml

[LudwigDevOps]

Hello Guys,

i have a simmular problem with this.
i have prometheus and snmp_exporter on the same machine (no docker, just debian OS)
prometheus runs on 9090 and snmp_exporter on 9116

in the prometheus config file i have declated my targets:

  - job_name: snmp
    metrics_path: /snmp
    params:
      module: [if_mib]
    static_configs:
      - targets:
        - 192.168.130.246 #HP ProCurve 4204
        - 192.168.130.252 #HP 2530
        - 192.168.130.249 #Aruba 6100
        - 192.168.130.250 #Dell PowerEdge 2824
        - 192.166.130.1   #Firewall 
    relabel_configs:
      - source_labels: [__address__]
        target_label: __param_target
      - source_labels: [__param_target]
        target_label: instance
      - target_label: __address__
        replacement: myserver.domain.com:9116 # URL as shown on the UI

the other switches are running and everything is fine.
i collect the data and display my traffic with grafana.

but the firewall one is a little bit bitchy.
the snmp config on the firewall should be fine snmp v2c and public as the com.string.

on the webinterface from the snmp_exporter the firewall(called devices) gives back his information instant.
webinterface runs with this URL:

http://myserver.domain.com:9116/snmp?target=192.168.130.1&module=if_mib

fine.. i got my ifOutOctets

on the prometheus webinterface on targets.. i got the error 500 and if i click on the link i get
error getting target 192.166.130.1: request timeout (after 3 retries)

i look in the URL and its this one:

http://myserver.domain.com:9116/snmp?module=if_mib&target=192.166.130.1

so just the structure of the URL is changed.. and then error 500 on prometheus and timeout????

can you guys say me something about this. why the URL stucture changes and why it workes with just this constellation..
it makes no sense to me.

br

Ludwig

btw: i use this version

 snmp_exporter --version
snmp_exporter, version 0.22.0 (branch: HEAD, revision: 7b0657fd99862a47f2f9f780f08a794ca7040ae7)
  build user:       root@cfac3556bd96
  build date:       20230615-13:29:47
  go version:       go1.20.5
  platform:         linux/amd64
  tags:             netgo