What do you use to monitor performance of your hosts and containers?
What do you use to monitor performance of your hosts and containers?
I have a Raspberry Pi running Docker and a number of containers. I plan on adding another Pi soon. Curious what folks are using for a dashboard to monitor performance of all your hosts and containers. I was thinking of deploying Grafana for this but am curious what others use.
Grafana, VictoriaMetrics (drop-in replacement for Prometheus with better storage efficiency and enhanced query language), Loki, Telegraf and Promtail for metrics and logs correspondingly.
Example with provisioned datasources and dashboards here:
https://gitlab.com/homelab_software/monitoring
It’s possible to ship metrics and logs with cAdvisor and other tools.
1Well, I run my containers in kubernetes.
And, it more or less includes full support for prometheus/grafana/alertmanager/etc.
So- I use that.
1I have questions about this. I’ll be getting another Pi or two and was considering putting k8s on them. Would I be able to set them up with kubernetes and then import my existing Docker containers from my current Pi to them?
1Yup. You can do that.
Although- you wouldn't "import" your existing containers. but, you can...
- Create manifests for your containers (Kubernetes runs the exact same docker containers). or, find helm charts for your containers.
- Import the storage from docker into your new PV/PVCs.
I would, suggest learning kubernetes first though. Learning curve can be rather steep.
Also, rancher + k3s would work perfect for your Pis.
1
Does overspeccing your hardware so much performance issues never come up count?
For normal people grafana & prometheus are typical good answers.
1Flag a warning when usage hits 5% so you can start saving for the next server
1
I'm definitely alone but... Checkmk. AMA.
1Telegraf with Docker Input Plugin installed on the Host writing to InfluxDB and displayed in Grafana, both running in Docker containers.
Here is a screenshot of my Server Performance Grafana dashboard.
1Netdata
1Does netdata support multiple servers? Can I see statistics for all hosts in a centralized way?
Thank you!1
vSphere lol
1Datadog.
1Zabbix and a TIG stack
1Munin. It’s highly portable and works with many operating systems.
1PRTG
1Only reason I keep a Windows box around!
1
Prometheus and Grafana for production environments
1I really enjoy practicing with Datadog - though it gets quite expensive really quickly and is quite overkill for 6-7 hosts, many VMs, and 20ish containers.
We use it at work, but monitoring isn’t my team’s responsibility so I try to understand how it all fits together by practicing with it at home.
I think Datadog should have a homelabber tier (above the free 5 physical hosts) that allows people to tinker. I honestly think it would net them more customers.
1Ahhh, Datadog, the sleazy used car salesmen of the observability market. Seriously, they're hucksters.
1
New Relic
1I’m curious which monitoring tool is the easiest to deploy and maintain. I’m looking to deploy a monitoring solutions via docker on an existing server. I wasn’t a fan of Zabbix with their docker deployment.
1Home Assistant.
1Netdata, works well for me
1Mostly my own eyes.. /s
I run Dozzle as a container on my host and I use the command 'docker stats' on the CLI on the dockerhost for in-depth stuff.
1Telegraf, InfluxDB and Grafana
1Zabbix
1I second this. I just set my instance up a couple weeks ago.
While researching, I saw many people saying that it’s very good but is hard to set up. I disagree with that to some degree; Setup itself is extremely easy, configuring it the way you want isn’t as easy BUT is wayyy more time consuming. Time consuming != hard, though. Just take time to tweak it how you want.
1What specs did you put it on?
1I second this. Setup was a breeze for me compared to checkmk.
1
