Ben Kochie

@lukas-mbag this seems to be abandoned long enough that I would go ahead with what you want to work on. You're welcome to cherry-pick changes from this PR if that...

This needs a rebase.

Closing as stale. Anyone seeing this is welcome to open a new PR with a rebase.

I've also experienced this with different clients. Multi-client functionality is handled better in OTRv3 protocol. (libotr 4.0) It looks like this implementation only supports OTRv2 and OTRv1

Yes, that very much seems like something about the device. Best you can do is maybe tcpdump/wireshark the scrape to compare send and receive times to look for a response...

We would need to have retry callback hooks in the SNMP library in order to get more per walk debugging. It may be possible, I would have to look at...

Can you give this PR a try? https://github.com/prometheus/snmp_exporter/pull/1157 You can find the output build artifacts in [CircleCI](https://app.circleci.com/pipelines/github/prometheus/snmp_exporter?branch=superq%2Fretry_debug).

I would look for `Retry number` and `request timeout` in the logs.

Yea, I typically don't recommend running SNMP over VPN/WAN links. SNMP, being UDP-based, is very sensitive to packet loss and latency. My standard recommendation is to run the snmp_exporter as...

We have talked about having Docker support in the past. What about podman support directly instead of systemd?