Hi guys,
Previously I've announced the es-diag tool -
https://github.com/jondot/es-diag, a readyness/healthcheck toolkit for
elastic search.
After reviewing parts of the mailing list and Github issues looking for
specific best practices, it is hard to find something that I can mark
off as an obvious pitfall or best-practice. It think, therefore that it is
best to start with the general opinion directly (welcome any other idea).
So, if anyone have best-practices, pitfalls, checklist or things they wrote
on a back of a napkin, such that they can share,
feel free to express it in words, and I'll take care of adding it into the
tool as proper 'checks'. You could then pick up the
latest version of the tool and run it for your own convenience.
The main things I'd need in a good description would be
- source of data (can be anything really)
- what are the conditions that are considered bad
- how to amend it, in general
Just to get a feel for how such a thing looks coded, I've already done one
example based on Shay's recommendation for JVM heap settings, out in 0.0.3
which you can view here:
The idea is to have a coded 'spec' of checks, that anyone can run, either
before setting up elastic search or while it is running
to make sure the server is always optimized to host it. Further, it may
bring troubleshooting and support to a no-brainer level.
Thanks!