Update haproxy.cfg to silence warning about 0 for server timeout

[esabol]

This PR partially addresses issue #123 by changing the configured value of the server timeout for the docker-events backend from 0 to 30s. It silences the following warning message:

[WARNING] While not properly invalid, you will certainly encounter various problems with such a configuration. To fix this, please ensure that all following timeouts are set to a non-zero value: 'client', 'connect', 'server'.

[pedrobaeza]

Isn't the amount very high?

[josep-tecnativa]

Maybe a better timeout would be something like 5 or 10 seconds

[esabol]

@pedrobaeza @josep-tecnativa : I'm happy to use whichever number you prefer. 30 is just what was suggested in issue #123.

[esabol]

It's been pointed out that the 0 value comes from PR #98. Maybe this shouldn't be changed?

[josep-tecnativa]

Thanks for sharing this PR. I’ve been investigating a bit, and the /events endpoint is a long-lived streaming connection. Docker keeps sending data as long as the client is listening, so the connection doesn’t really hang. That’s why a timeout is not needed here — setting one would only risk breaking valid event streams. I see that the 0 value comes from PR #98
, so maybe it was intentional for this reason. In any case, the warning fix doesn’t really seem to come from changing the timeout — or at least it’s not a valid solution for us, since it could cause issues.

I'm closing this PR. Thank you anyway.

[esabol]

In any case, the warning fix doesn’t really seem to come from changing the timeout

Changing the timeout definitely does eliminate the warning message. I tested that. That said, I agree that it shouldn't be changed just to eliminate the warning message for the reason given in PR #98.