[Labs-l] [Wikitech-l] Brief Labs outage

Fri Aug 1 06:48:08 UTC 2014

On 01/08/14 08:21, Petr Bena wrote:
> problem of these 2 is that they are maintained primarily by puppet,
> which for good reasons is loathed by many and disabled or killed on
> many instances, which result in these monitoring tools being
> defunct there.

Hi,

could you please elaborate on "which for good reasons is loathed by many"?

I'd like to understand if the reasons are:

- some inherent flaw of our labs puppet structure
- general dislike for configuration management software
- a specific dislike for puppet

because in the first case we can work to fix the flaws, and that would
solve the problem.

 I don't honestly think there is a way to have monitoring without some
config management system collecting checks on the single labs
instances, so we should think of making puppet viable for most (at
least for the base system) if we want to have monitoring in labs.

Cheers,

Giuseppe

-- 
Giuseppe Lavagetto
Wikimedia Foundation - TechOps Team