Ask HN: What are your monitoring data retention policies, and why?

3 points by movedx ↗ HN
I'm in the process of putting in place a monitoring solution which, to begin with, will collection information on:

* CPU, RAM, disk, and networking; * The system's load (averages, # of procs, etc); * User information (logged in, and so); * HAProxy, MariaDB, MongoDB, ElasticSearch;

I'm wondering what a good retention policy would be for a generic monitoring solution like this? And Why?

What policies does your business mandate and what is the reasoning behind them?

1 comment

[ 4.1 ms ] story [ 89.1 ms ] thread
We adjust it so that it uses 80% of the monitoring servers disk.

Just store as much as you can - "some" metrics is more valuable than an empty disk, even if only marginally, and you never know how you might want to slice and dice the data in future.