edgard / iperf3_exporter

Simple server that probes iPerf3 endpoints and exports results via HTTP for Prometheus consumption
Creative Commons Zero v1.0 Universal
66 stars 26 forks source link

Got kill signal with, "level=error msg="Failed to run iperf3: signal: killed"" #18

Open mrezabarkhordari opened 1 year ago

mrezabarkhordari commented 1 year ago

Hi, after running iperf3_exporter from time to time and randomly (without a specific pattern) got this error which tells the iperf3 is killed do you have any idea?

Benedict1033 commented 5 months ago

Have you solved the problem? I have the same problem

fpmb commented 2 weeks ago

Realise this is old but just ran into the same or similar...

Error message could perhaps be improved (apologies for being lazy and not submitting a PR). For me, it was iperf3 timing out as I had the scrape_interval in Prometheus set low for testing purposes which caused the scrape_timeout to kick in on my slow network.

The scrape_timeout for a scrape job (or globally) has to be less than the scrape_interval which makes sense... don't want Prometheus repeating the same testing when the previous one hasn't completed.

Example:

  - job_name: 'iperf3_exporter'
    scrape_interval: 5m
    scrape_timeout: 1m
    [...]

... so scrape_interval must be infrequent enough to allow iperf3 to complete a test which for me meant scrape_timeout of 1m was plenty of time for the iperf3 test to complete even on my slow network :)

Many thanks to @edgard for the useful image!