Feature #12161

Monitor that our internal XMPP server is up

Added by intrigeri 2017-01-21 17:42:40 . Updated 2018-01-18 14:59:59 .

Status:
Rejected
Priority:
Normal
Assignee:
Category:
Infrastructure
Target version:
Start date:
2017-01-21
Due date:
% Done:

0%

Feature Branch:
Type of work:
Sysadmin
Blueprint:

Starter:
Affected tool:
Deliverable for:

Description


Subtasks


Related issues

Blocked by Tails - Feature #12162: Give the internal XMPP service admins the credentials they need Resolved 2017-01-21

History

#2 Updated by intrigeri 2017-03-08 15:16:55

  • blocked by Feature #12162: Give the internal XMPP service admins the credentials they need added

#3 Updated by intrigeri 2017-03-08 15:17:36

  • Target version changed from Tails_2.12 to Tails_3.1

I’ll take it easy too.

#4 Updated by intrigeri 2017-06-01 14:36:25

  • Target version deleted (Tails_3.1)

I’ll re-add a target version once there’s something to monitor.

#5 Updated by intrigeri 2017-06-05 17:21:46

  • Assignee deleted (intrigeri)

intrigeri wrote:
> I’ll re-add a target version once there’s something to monitor.

Ditto for the assignee.

#6 Updated by intrigeri 2017-07-25 09:30:56

  • Assignee set to intrigeri
  • Target version set to Tails_3.2

It’s up!

#7 Updated by intrigeri 2017-07-29 17:42:46

The only candidate plugin in Debian is https://tracker.debian.org/pkg/nagios-check-xmppng:

  • homepage: https://exchange.icinga.com/jandd/check_xmppng
  • last upstream release: 2016-06-18
  • very low popcon but that’s somewhat expected
  • no support for SOCKS5 so we would need to wrap it somehow with torsocks (I think we do that already for SMTP → WhisperBack relay)
  • no reply on https://bugs.debian.org/846873 for almost 8 months; we’re not affected by it but that’s somewhat concerning wrt. the maintenance of the Debian package; now, that’s not any worse than importing 3rd-party plugins from the Internet into our own Git repo without any process set up to update them, so well, I think we can live with that.
  • in jessie-backports, Stretch, and testing/sid

#8 Updated by intrigeri 2017-07-30 13:34:25

  • Target version changed from Tails_3.2 to Tails_3.5

#9 Updated by intrigeri 2017-12-10 08:35:37

  • Assignee changed from intrigeri to Dr_Whax
  • QA Check set to Info Needed

Taking a step back, I’m now in doubt. I wonder what’s the actual value of implementing this:

  • This is purely internal infrastructure, if it’s down only almost-core Tails contributors are affected; the dev process and users are not.
  • A few of us are on tails-bar all the time so chances are we’ll notice if this service is down as fast as a monitoring check would.

So the only case when doing this work would be useful is when a monitoring check would report a problem to the on-call sysadmin, and the on-call sysadmin fixes it, all this before any tails-bar user notices the problem. Given our average latency for reacting to such non-critical issues raised by our monitoring system, I doubt this ever happens.

So I propose we reject this ticket and encourage our sysadmins to hang out on tails-bar when they’re on duty (to increase the chances they notice any issue at the same time as other users of the service).

DrWhax, bertagaz, groente: what do you think?

#10 Updated by Dr_Whax 2018-01-18 13:39:47

intrigeri wrote:
> Taking a step back, I’m now in doubt. I wonder what’s the actual value of implementing this:
>
> * This is purely internal infrastructure, if it’s down only almost-core Tails contributors are affected; the dev process and users are not.
> * A few of us are on tails-bar all the time so chances are we’ll notice if this service is down as fast as a monitoring check would.
>
> So the only case when doing this work would be useful is when a monitoring check would report a problem to the on-call sysadmin, and the on-call sysadmin fixes it, all this before any tails-bar user notices the problem. Given our average latency for reacting to such non-critical issues raised by our monitoring system, I doubt this ever happens.
>
> So I propose we reject this ticket and encourage our sysadmins to hang out on tails-bar when they’re on duty (to increase the chances they notice any issue at the same time as other users of the service).
>
> DrWhax, bertagaz, groente: what do you think?

Yeah, I think that makes sense, also, tails-bar is nice, so please hang out there :)

#11 Updated by intrigeri 2018-01-18 14:59:59

  • Status changed from Confirmed to Rejected
  • Assignee deleted (Dr_Whax)
  • QA Check deleted (Info Needed)

Rejecting then. If someone feels differently, feel free to reopen (and either explain why it’s important enough for me to do the work, or assign the ticket to yourself :)