Kubernetes 101 - Episode 10 - Monitoring with Lens, Prometheus, and Grafana
ฝัง
- เผยแพร่เมื่อ 2 ส.ค. 2024
- In this episode, we’ll talk about different ways of monitoring Kubernetes, including Lens for better cluster visibility, or tools for metrics, like Prometheus and Grafana.
More info: kube101.jeffgeerling.com
Special thanks to amazee.io for sponsoring this series: www.amazee.io
Sponsor me on GitHub: github.com/sponsors/geerlingguy
Support me on Patreon: / geerlingguy
Links:
- Kubernetes 101 Book: www.kubernetes101book.com
- Ansible for Kubernetes for $4.99: leanpub.com/ansible-for-kuber...
- Episode Details: Coming soon!
- Episode Examples: github.com/geerlingguy/kubern...
#kubernetes #kube101
Contents:
00:00 - Livestream start
00:50 - Welcome!
02:04 - Monitoring Kubernetes
06:28 - Cluster Visibility with Lens
11:19 - Exploring a Cluster with Lens
22:00 - Prometheus and Grafana
26:40 - Install the kube-prometheus-stack
36:59 - Monitor your cluster with Grafana
41:49 - Final thoughts on monitoring
44:07 - Conclusion + Kubernetes 101 Book - วิทยาศาสตร์และเทคโนโลยี
Be sure to check out the book based on this series! www.kubernetes101book.com
Thanks so much for watching, to those of you who made it through all ten episodes :)
This is how I view all discussions of K10 th-cam.com/video/i3RqDp7i308/w-d-xo.html
Doesn't Lens also deploy it's own resources into the cluster ?
In terms of watching logs form a bunch of pods. Take a look at github.com/johanhaleby/kubetail
@@DontScareTheFish Nope! If it did, I wouldn't touch it with a 100' pole.
It does have some options that you could 'inject' like a custom Prometheus configuration, but that's not on by default, and if they ever tried putting that in by default I'd turn against Lens in a heartbeat.
Bought the book, looking forward to seeing it grow. Thank you for sharing your knowledge!
Now that's what I call a great series. Tons of useful information presented in a pleasant and calm manner. Kudos!
Hey Jeff, thanks for giving to the community. Amazing series. God bless.
Great series! Thank you for making it. And the operators one was my favorite episode!
I must say that Operators episode was the eye-opening moment for me. It could be that the topic is more advanced than what is expected from a 101 series. I really hope there will be more intermediate content
Fantastic series, Jeff!
Even having worked with k8s for a few years now, I still learned some new things and cemented a few concepts that were still nebulous!
Awesome series! Thanks a million!
Thank you for the detailed explanation. Your videos are the best 👍 I have no need for a cluster but this has helped me scope the infrastructure/tools I could use in the future 🚀
Awesome K8s monitoring intro!! 👏👏 I am looking forward to learn a bit more on the alerting topic with these tools 👀
This is a great presentation! Keep the good works!
Great job Jeff. Again. Thank you!
Excellent video! Just what I was looking for!
Thanks you for your videos! Continue and do not give up.
Dear Jeff, a lot of thanks for this video.
Hi Jeff, thank you for this series - please finish your book :) really enjoy your videos
Top notch video! 🚀
Thank you very much, Jeff! You are very cool and amazing man.
I have just bought your book. I hope it is a good as the Ansible for Devops! Best regards from Germany!
Cool video! Thx! Tell please about logging (collecting, analyzing, alerting) in k8s in next your video.
Thanks for the book, But on a point in time i wlll bing All the Kubenetes 101 Episodes
Another excellent episode. Thanks a lot Jeff!
I've tried deploying Prometheus and Grafana using the helm chart you mentioned on my Raspberry pi K3S cluster and I've got the prometheus-kube-state-metrics pod on CrashLoopBackOff. It's the only pod that's not running.
Logs said: standard_init_linux.go:219: exec user process caused: exec format error │ stream closed
I've assumed the chart is only AMD64 compatible. I would be really great if more of these tools were multi arch.
Next step, I will try it in my other AMD64 K8S cluster.
Till next time.
If you're on Pi / ARM, please give carlosdp's cluster-monitoring stack a try: github.com/carlosedp/cluster-monitoring
It's the same thing, but uses images guaranteed to work on ARM/ARM64.
@@JeffGeerling I will try it on my Raspberry K3S Cluster later. Thanks for the suggestion.
Don't you think that cross platform apps compiling and OCI cross platform image building would be a great topics for future episodes?. What are the options apart from Docker build and buildx?.
I will really enjoy to learn how to do that. And it will help develop Raspberry Pi and ARM ecosystems.
Just thinking out loud!.
@@apulacheyt I can recommend Saint Jeff's 😉 book Ansible for Kubernetes. I seem to remember that targeted builds were in that book. I got that with his Ansible for DevOps book and it's a good combination. Gets my recommendation! 🤩
Very informative. What are your thoughts on rancher when compared to lens?
Hey Jeff, I really enjoyed thid serie!Het zou geweldig zijn als je een kubernetes 102 serie zou maken. Or in English, it would be great if you would make a 102 serie about kubernetes! Groeten uit Nederland!
Should we use Rancher and Lens IDE together or using them separately??
Some opinions:
The operators episode could have been part of a 102 series yes.
A Q&A section was REALLY needed. k8s is complicated, and a Q&A would cover some of the possible holes that people had.
Diving more into deploying & customizing stuff with Helm would have been appreciated.
Good points, I think a Helm deep dive episode might've been better than the Operators video. Also, I'm considering doing a separate 'Kubernetes Questions, Answered' video at some point, maybe a few videos over time. There are definitely some good ones I haven't answered from live chat and comments!
@@JeffGeerling I think the best way to describe my thoughts on the series would be to say that it crossed between a 101 and a 103. It also would have been nice to have a consistent format on discussion of a topic or problem and then implementing the topic/solution. You discussed logging problems/options in 6 for example, but did not deploy them. 1-3 did a great job of this. some of the others as well. Episode 10 centered around the insights problem at a high level, but that was not how it was approached.
k8s is complicated and there is VERY little good, *consistent* information on the topic. This series would be a great hand hold, but it still has its issues.
Hi, from Bannockburn, Australia. Not live because of the timezone.
I just started learning kubernetes and prometheus. I haven't used Lens before but it looks a lot like Minikube dashboard. Is it the same?
Man my autism picked a bad thing to wanna go for... Autism hates abstract concepts and yet I love IT. Even with a fancy UI I struggle to wrap my brain around all the concepts... I seem to only do well in school (BSIT/4 year degree) but when I want to do something nice for myself to monitor my own application server it just goes way over my head.
i am from Nepal 😃
great content, i would recommend to monetize your videos to get some ads so you can get something in reward ... thanks for these videos
How Can I define the period of store logs per container in Lens tool? I'm not able to chech logs from 6 h back? ???
microk8s status -wait-ready
microk8s is running
high-availability: yes
datastore master nodes: ...
Thank you!
#fromQuebec
which browser are you using? is that Brave browser?
Nope, just good ol' Safari.
Safari looks nice 🙂
How to make Prometheus storage persistent for 1 year
Please refrain from using the bogus term "Open Source". What you are talking about here is actually "Free Software".
I am in a situation where the node_exporter in always in pending state. And the port 9100 is already occupied tcp6 0 0 :::9100 :::* LISTEN 516561/node_exporte ; After doing kill -9, it get restarted itself ( kube-scheduler). Is there a way to get rid of this & get prometheus node exporter in running state.