[ProgClub programming] Centralised logging

Jedd Rashbrooke jedd.rashbrooke at gmail.com
Wed Jun 12 19:11:53 AEST 2019

On Wed, 12 Jun 2019 at 17:17, John Elliot V | ProgClub <jj5 at progclub.org>

> I have no experience with containers, so in addition to learning about
LVM I can use this as an opportunity to learn about containers, too.

 Then build your monitoring platform first, and then later play with
containers and log collectors.  Monitoring gives you visibility into the
performance of itself, as well as your other systems, so it's the most
important gap to fill.

So that brings us back to the disk requirements for my new VM. Still
> just start with 30 GB for the system? Or maybe bump that up to 50 GB?
> What are the suggested initial CPU and RAM settings?

 30GB for root, put everything on the one partition, put swapfiles in
rather than swap partition, refer my earlier messages detailing my block
layout, it's a good place to start.

 2 core, 1GB RAM.    On the assumption this is a headless box, and you
won't be installing a GUI on it.  The nature of VMs are that you can always
power them down, change core, memory, and disk allocation, so don't fret
about this.

 I'd suggest Prometheus would be easier and more useful than Zabbix.  On my
system I've got VSS of around 3GB, but RSS of less than 100MB, and I'm
using ~ 5GB disk after some months, with ~12 servers, and polling every 60s
with vanilla node-exporter metrics exposed on those hosts.

 Apart from prometheus.yml file for static config targets, I also changed
the systemd service file as follows, as the retention rate was too short by

 # ExecStart=/usr/local/bin/prometheus
 --storage.tsdb.path="/var/lib/prometheus/" --storage.tsdb.retention=730d

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.progclub.org/pipermail/programming/attachments/20190612/7f927ffe/attachment.html>

More information about the programming mailing list