Probably it is very incomplete. I copied over much of the content from 0.7 that is still applicable, but I added the following: 1. A section on data centers and mirroring 2. Removed the cut-and-pasted scala code in the monitoring section and just gave a high-level description of what we monitor 3. Added a lot of details about what we have learned about Linux, ext3, write latency, etc.
My thought was that it was kind of a pain to keep a few hundred metric names up-to-date. So we tend to end up with wrong docs. I wonder if it isn't easier just to let people fire up jconsole or whatever and browse through the full list of metrics that way?
-Jay On Wed, Jul 24, 2013 at 8:06 AM, Otis Gospodnetic < [EMAIL PROTECTED]> wrote: