Created
June 28, 2017 17:22
-
-
Save clayg/e87e2b9c69abbe25f45cc09d570687ec to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
> Should be that "handoff" and "misplaced" columns should ideally be 0 | |
not necessarily | |
> but to also perform the correct operations that need to be done in response to what it finds? | |
No, I don't think that's a good idea - a) it pollutes the role of metrics | |
collection and monitoring with alerting and intervention b) it's the job of the | |
consistency engine, if a silver bullet existed; it should go there. | |
> based upon some default metric or command-line defined metric | |
I don't know what *specifically* you're suggesting. This sounds like "why | |
can't it automagically just DO THE RIGHT THING" - which is a surprisingly | |
nebulous concept when you consider the immense variety of how swift is deployed | |
and used. | |
It's like the adage of raising toddlers: | |
Every story about dealing with toddlers goes the same, "we tried this, then we | |
tried this, then we tried this - and FINALLY we found something that works!" | |
When you meet parents that have only one kid they like to tell you about the | |
RIGHT THING TO DO; when you meet parents with multiple kids they like to tell | |
you "everyone is different, don't be afraid to try different things, you have | |
to learn what works for you" | |
FWIW, now that you're monitoring overall cluster health - you can start to | |
experiment with tuning up replicator concurrency and monitoring network | |
traffic, cpu, memory, disk io - and see what "works for you" |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment