Error database requirements

[Robert Collins]

On Fri, Aug 12, 2011 at 6:52 AM, Bryce Harrington <bryce at canonical.com> wrote:
> 1.  Prior art in Linux?  Aside from Breakpad, do any other distros or
>    FOSS projects have crash reporting systems like this?

There is LP's OOPS system, there is a google project for cross
platform crashdump capturing, a django project called 'sentry' for web
server error analysis (that has a cassandra backend I'm told) ...  I
suspect a plethora of small scale automated-gathering tools out there,
but the only massive scale end-to-end one I know of is breakpad.

> 2.  I notice the Windows crash reporter includes buttons to check for
>    existing solutions.  Is that something we'd like to consider as a
>    requirement for this?

Yes, but I think that should be done in future iterations - theres
plenty of meat around just handling:
 - file anonymously
 - without uploading too much
 - or too little
 - handling offline situations
 - data storage and backend scaling
 - initial analysis facilities
 - dealing with privacy
 - allowing users to follow up on their report [if desired - note that
Microsoft's tool *used* to do this, now they don't]

> 3.  Privacy issues should probably be elaborated on further.
>
> 4.  How do us developers want to interact with the gathered data?
>    I.e., what should the graphs or tables look like?

I suggest in a first pass we want a strong API - probably a map-reduce
style in-the-cloud processing workflow, and use that to build a real
web UI as well as adhoc queries and custom analysis.

-Rob