[okfn-help] OKFN Dashboard

Tue May 31 08:29:53 UTC 2011

On 31 May 2011 01:09, Tim McNamara <paperless at timmcnamara.co.nz> wrote:
> On 31 May 2011 04:06, Rufus Pollock <rufus.pollock at okfn.org> wrote:
>> Hi Tim,
>>
>> I see you started a new repo but then discovered we already have a repo here:
>>
>> <https://bitbucket.org/okfn/okfn-dashboard>
>
> :)
>
> Did you see this comment? http://trac.okfn.org/ticket/713#comment:7

Yes, hence comment about discovering the other repo ;)

> The work that I did last night was mainly just testing out watchdog.
> Polling 40+ sites seemed like the easiest way to start.

Is this for uptime etc -- at the moment we are planning to use nagios
for that (and already have it set up -- Nils dealt with this).

> As it happens, watchdog is excellent. Now that I've seen an
> alternative, I feel that cron is not ideal. Manual tweaks would need
> to happen every time a service was added or the monitor was moved to a
> different host. I would like to be able to make sure that the
> monitoring system can be run anywhere.

Looking at watchdog it seems largely to be about the kind of 'up-time'
monitoring that we already have covered (to a reasonable extent).
While I'm not specifically objecting to looking at other alternatives
(or adding requirements like: "should run anywhere") I think this is
definitely lower priority than the main work on the dashboard (which
is a 'community dashboard' not a 'sysadmin status dashboard'). What do
you think?

>> As there are several people who have offered to contribute I suggest
>> it would be good to try to finalize planning and parcel up the work a
>> bit focusing on our existing ticket (and any sub-tickets we can
>> create):
>
> +1
>
> Is this mailing list the right forum for that discussion?

Yes, I think so.

>> A starting point would be to get a really clear set of use cases (see
>> the ones i've already created). We can also avoid duplicating anything
>> we've done already (e.g. Nils our sysadmin has already got a list of
>> sites and services and nagios monitoring set up at status.okfn.org).
>
> Some cases to think about:
>
>  - an outsider looking to see what OKFN is about
>  - having something to show off to people at a monitor
>
> Are any of these use cases that should be included

Yes, agree on both though we probably want to try and flesh these out
so we have something very specific to start on.

>> I think there are probably a bunch of separate tasks:
>>
>> * Getting a list of sources
>> * Pulling / scraping data from those sources
>> * Integrating that data
>> * Exporting it as reports and providing front-end web pages
>> * We may want an overall mini-webapp though i think, at the start, we
>> should aim for just json plus flat html as much as possible :-)
>
> +1
>
> I don't see any reason to provide anything more than static HTML.  If
> admins want access to the database, they can use SSH.

Great, agreed.

Rufus