Alertmanager: env http_proxy stopped working

Created on 24 Mar 2018 · 3Comments · Source: prometheus/alertmanager

What did you do?

Upgrade Alertmanager from version to 0.12.0 to 0.15.0-rc.0.

What did you expect to see?

In my previous configuration I exported the environment variable (http_proxy) to send alerts. It doesn't work in latest version, there is a new option global/http_config/proxy_url which does the same job.

Can you confirm it ? For me it was the breaking change so it's worth to mention about in release notes,

Thanks

Source

tomplus

👍3

Most helpful comment

Sure, but a small notice costs nothing but can save somebody's time.

tomplus on 24 Mar 2018

👍7

All 3 comments

You can now configure proxies from the configuration file. The environment was never part of our API.

brian-brazil on 24 Mar 2018

Sure, but a small notice costs nothing but can save somebody's time.

tomplus on 24 Mar 2018

👍7

@brian-brazil related: does the HTTPS_PROXY have an analogue in the configuration file? As of now, setting global/http_config/proxy_url doesn't seem to set it correctly, for OpsGenie at least:

Apr 13 17:38:57 hostnamehere alertmanager[66535]: level=error ts=2018-04-13T21:38:57.337971224Z caller=dispatch.go:266 component=dispatcher msg="Notify for alerts failed" num_alerts=1 err="Post https://api.opsgenie.com/v2/alerts: proxyconnect tcp: tls: oversized record received with length 20527"

I can open a new issue if this is unexpected or broken behavior. Currently it makes OpsGenie unusable through a proxy for us.