fluentd is resource hungry. The CPU usage slowly kept rising over 8-9 hours and suddenly spiked the memory from 1G to 8G crashing the nginx server itself causing downtime in production. This happened when the nginx server was handling just 5-6RPS.
I just figured out you have almost everything. Do you have a human brain or a chip 🤔 I was also wondering if you offer training in python programming for DevOps engineers
assuming that it's because the program won't have permissions to make changes to my files that have permissions set so that other users can't open them...but they can still read my files. I guess there's probably some nuance here I'm missing that makes it worth it.
@@samtx You could use Docker Swarm, but it has been outdated for a while. If you need container orchestration (running containers on multiple servers), consider using k8s (Kubernetes).
Hi @Anton Putra, Could you please make a similar video on k8s cluster to monitor Nginx with Prometheus and Grafana using fluentd.! It would be of great help.
@Anton Putra Nice video content. Followed through but noticed at the end that the imported fluentd dashboard failed with the error `Templating Failed to upgrade legacy queries` and was not able to display anything. Any idea what could be the problem as other dashboard i imported via ID worked fine or is there a way to convert from JSON to ID?
Which capability Fluntbit does not have? Can you bit more clear on it please. Fluentbit which run as daemon set, hence should be able to read parse the log stream of Nginx?
I know, I just couldn't find the 3rd party plugin that i can use to parse logs and expose them as prometheus metrics. If you know one, please let me know.
Amazing video, thank you! I have the following error while importing the 2nd group of panels: "Failed to upgrade legacy queries Datasource r_9gjRp4z was not found" how can i fix it? thanks
Very nice tutorial, thanks 👍 Side note: I'm just wondering why you don't use Ansible or any other CM (config management) tool to install and configure all the dependencies ?!?
@@AntonPutra , а какую именно переменную ты имеешь в виду? value в json-е на гитхабе со значением /api/devices? Помоги пожалуйста, уже долго пытаюсь настроить мониторинг 4xx и 5xx ответов nginx-а в grafana, но пока без успешно.
❤Part 2 - How to Build Custom Prometheus Exporter? - th-cam.com/video/3wT0zSsQb58/w-d-xo.html I misspoke; this particular nginx prometheus exporter must be installed on the same VM compared to the kafka exporter, where you can specify multiple brokers.
🔴 - To support my channel, I’d like to offer Mentorship/On-the-Job Support/Consulting - me@antonputra.com
Excellent tutorial and I have implemented as it is and worked like charm. Keep going bro.
❤️
Thanks. Your posts videos are very useful when trying other monitoring tools and approaches!
Another great video. Thanks Anton
My pleasure!
Thanks Great Video as always, lot of learning
Thanks Muhammad!
Today is my first day at close-to-devops job and this video is helpfull.
It's a very good video, very clean and organized.
Thanks Ferdinando!
Thanks. Your posts videos are very useful when trying other monitoring tools and approaches!
Glad it was helpful!
fluentd is resource hungry.
The CPU usage slowly kept rising over 8-9 hours and suddenly spiked the memory from 1G to 8G crashing the nginx server itself causing downtime in production.
This happened when the nginx server was handling just 5-6RPS.
Agree, try lua plugin - github.com/knyar/nginx-lua-prometheus
Let me check it out, thanks
Thank a lot, for this guide!
my pleasure!
Thanks Anton
welcome!
Keep it up bro !!!
Thanks! Will do
Great video @anton
Can you make a video on how to set up this solution on EKS?
Is there anything specific? Cause I have bunch of them on my channel if you search for Prometheus or Eks
I just figured out you have almost everything. Do you have a human brain or a chip 🤔
I was also wondering if you offer training in python programming for DevOps engineers
do you have any videos with explainer on why you use a different user for each process?
In general you want to use a different user for each process for security reasons.
assuming that it's because the program won't have permissions to make changes to my files that have permissions set so that other users can't open them...but they can still read my files. I guess there's probably some nuance here I'm missing that makes it worth it.
awesome
Thanks!
Great video thanks a lot
Thanks!
hi your dashboard no this in the github repository
perhaps forgot to commit, but it's simple enough to recreate
using docker-compose would have been a massive time and complexity saver in this project.
sure, i have an example in the following video - github.com/antonputra/tutorials/blob/main/lessons/141/docker-compose.yaml
@@AntonPutrahow to run docker compose on two nodes
@@samtx you can use minikube and spin up multiple nodes
@@AntonPutra thx I was asking if docker swarm is need if to run dockercompose on multiple servers for hardware issues
@@samtx You could use Docker Swarm, but it has been outdated for a while. If you need container orchestration (running containers on multiple servers), consider using k8s (Kubernetes).
Hello, can u give a dashboard example for "latency traffic" like in this video??
*Sorry for my bad english
sure, here is source - github.com/antonputra/tutorials/tree/main/lessons/140/dashboards
just import it
I find it hard to understand your pwd as it's not shown in the terminal :(
Otherwise awesome material thanks Anton!
Thanks for the feedback, I'll include them in the readme starting next video.
Hi. Can you explain how can we setup the terraform locally without the AWS please?
Thak you.
You can use "docker-compose up" instead. Terraform is to create AWS VPC and EKS,
Возможна высокая кардинальность метрик по path
Soglasen, but it's a good starting point
Hi @Anton Putra,
Could you please make a similar video on k8s cluster to monitor Nginx with Prometheus and Grafana using fluentd.! It would be of great help.
how to install set-alert? is the method the same as the Prometheus set-alert in general?
do you mean alertmanager? yes it's a binary, you can use systemd to manage it as well
@@AntonPutra yes, that's what i mean
thanks you
@@fazlurrahmatullah3017 you can scroll down antonputra.com/monitoring/install-prometheus-and-grafana-on-ubuntu/
@Anton Putra Nice video content. Followed through but noticed at the end that the imported fluentd dashboard failed with the error `Templating Failed to upgrade legacy queries` and was not able to display anything. Any idea what could be the problem as other dashboard i imported via ID worked fine or is there a way to convert from JSON to ID?
try to update datasource or just create a new dashboard and import quiries
@@AntonPutra thanks got it fixed already🙏
@@samuelude6858 Heyy! can you please say how you solve this? i got same error
Great!!!
Which capability Fluntbit does not have? Can you bit more clear on it please.
Fluentbit which run as daemon set, hence should be able to read parse the log stream of Nginx?
I know, I just couldn't find the 3rd party plugin that i can use to parse logs and expose them as prometheus metrics. If you know one, please let me know.
and what about monitoring nginx unit with prometheus and grafana?
Unit? I haven't used NGINX Unit directly, mostly ingress.
Amazing video, thank you!
I have the following error while importing the 2nd group of panels:
"Failed to upgrade legacy queries Datasource r_9gjRp4z was not found"
how can i fix it? thanks
yes, can you try to update data source under settings (variables)
@@AntonPutraWhat exactly to update, seems it's the same issue with mine.
Great video, can you make a video how to do something similar with nginx-ingress in K8s? I think it will be very useful
Well, i have bunch of videos to monit nginx ingress including this one - th-cam.com/video/VjFFzGFyVlY/w-d-xo.html
Very nice tutorial, thanks 👍
Side note: I'm just wondering why you don't use Ansible or any other CM (config management) tool to install and configure all the dependencies ?!?
Thanks, I do, maybe for other videos - github.com/antonputra/tutorials/tree/main/lessons/151/ansible
Здравствуй, а как настроить для всего nginx-са , а не только для твоего тестового flask приложения?
Well, it's not only for the flask. You can use the path as a variable to switch between different upstream servers.
@@AntonPutra , а какую именно переменную ты имеешь в виду? value в json-е на гитхабе со значением /api/devices?
Помоги пожалуйста, уже долго пытаюсь настроить мониторинг 4xx и 5xx ответов nginx-а в grafana, но пока без успешно.
❤Part 2 - How to Build Custom Prometheus Exporter? - th-cam.com/video/3wT0zSsQb58/w-d-xo.html
I misspoke; this particular nginx prometheus exporter must be installed on the same VM compared to the kafka exporter, where you can specify multiple brokers.
So fast as usual, hard to follow
I kinda like the speed, gets straight to it
Maybe try playback speed of 0.75 I tried it and seems like could help if you find it too fast
You can follow the blogpost - antonputra.com/monitoring/monitor-nginx-with-prometheus/#intro
Templating
Failed to upgrade legacy queries
this is what i get in grafana. 2024 october. maybe ur version of grafana is younger than mine