launchpad-dev team mailing list archive

[Martin Pool <mbp@xxxxxxxxxxxxx>]

2009/8/14 Matthew Revell <matthew.revell@xxxxxxxxxxxxx>:
> I think that's the hard part. A manually updated system status page
> could quickly become inaccurate. Can't we automate it somehow?

Automation is very nice, but this spec has been waiting a long time,
and there have been multiple unexpected critical outages in the
interim.  They confuse users and make the system look more flaky than
it actually is.

What I'm suggesting is a page that describes only known critical outages.

We don't need automation to add things to that page because by
definition if they're known, they have someone's attention.  We just
need to add this small step of updating the page to the critical
outage procedure.

Similarly, when we either fix or downgrade the issue, we just need to
update it to say so.

> I wonder, also, if we might want to host it on a server completely
> unrelated to LP. ISPs and web hosts often have an off-site status page
> because, in some cases, what's taking everything else down may affect
> the status page.

That's a good point.  Robert raised the same issue on irc.  A
wordpress blog might be ok, for example.  Or a google docs page.

-- 
Martin <http://launchpad.net/~mbp/>