Grafana Dashboard📊: Monitor CPU, Memory, Disk and Network Traffic Using Prometheus and Node Exporter

[@tech_with_moss]

TOC: Introduction: 00:00 - 1:44 CPU metric: 1:45 - 09:03 Memory Usage: 09:04 - 14:15 Disk Usage: 14:16 - 21:20 Network Traffic: 21:21 - 25:06 Conclusion: 25:07 - 26:02

[@kshitijapatil6591]

is it possible to write cpu and memory usage one single query for one namespace?

[@mrloc274]

thank Moss, your tut very useful

[@mingyihorusli2068]

looking forward to more Grafana tutorials, thx

[@gourav09071994]

I am using Linkerd Service mesh managed Grafana..while I can see the pods, container, and cluster cpu and memory utilizations, I am not getting the system services metrics as the dashboard shows no data. I am not sure why. How can I get it through query or any ways I can resolve this pls?

[@moeburn]

bruh this was probably one of the best tutorial videos I have ever seen on KZbin. I got my Prometheus and Node Red and Grafana set up but I was like "how do I graph them?" and then your video comes up. So I do the first step and I'm like "ew but I have 8 cores", and then your video is like "so here's how to average all cores". And then I'm like "ew but these labels suck" and then your video is like "here's how to clean up the labels". And the best part, was that I wanted an average of all modes, just one CPU usage count not idle and system and user etc. But by watching you do an average of all cores, I was able to learn and figure out how to do an average of all modes too on my own! I learned something from your video! So awesome so helpful thank you very much.

[@tech_with_moss]

Thanks a lot. I appreciate it and I'm glad the video was so helpful!

[@RadhamaniRamadoss]

This is very helpful..Thankyou for taking time to share this excellent video tutorial.

[@janiwassim1836]

as in intern whose life depends on grafana...you are my superman

[@tech_with_moss]

Thank you! I'm glad the video helped

[@yousefirman1523]

this is a great video. I have searched some hours and this is only useful reference for alerting cpu/mem in Grafana, that I have found.

[@tech_with_moss]

Thank you so much for your comment. I'm really happy to hear that after a couple years or so of being posted that the contents of the video are still relevant.

[@mohammadshaazad4131]

I am so impressed by the way of explaining everything so that everyone can understand it easily! Thank you so much for this video.

[@tech_with_moss]

Thank you for the kind words. I appreciate it and I'm glad that my videos help!

[@ravism1585]

Really useful, I was struggling like hell to create dashboard. Thanks a ton

[@tech_with_moss]

No problem! Glad it helped.

[@mirnawazali322]

This video has been a life saver! Thanks for putting this out!

[@tech_with_moss]

No problem! I'm glad you found it valuable.

[@pesfreak18]

Very good video. Helped me a lot to make my first Dashboard in Grafana. Keep up with the good work.

[@abijithpradeep7726]

This helped me for my internship. Thanks a lot

[@bilaljaved6601]

Thanks For the Tutorial, it was really helpful for getting started with grafana.

[@rogervanerp8549]

Great video, I really appreciated the examples that you offered, I am new to Grafana and was wondering if there was a video or options to monitor Windows Core Servers clients with Prometheus and Grafana ? Thanks

[@Meowbay]

Can you also do per LAN IP network traffic to/fro internet with this setup? That would be grand, since we have switches on the interfaces..

[@carlomagnogonzalezvillalob3255]

Excelente video, muchas gracias por compartir

[@sujaysingh4299]

Thanks a lot for this Tutorial, You have explained it very nicely. Thanks once again!

[@КристианИлиев-х5щ]

The best video so far! Thanks.

[@yaiseravilarodriguez9226]

Awesome video, thanks for preparing it!! for the date of this comment, axes to control the units is located in Standard Options (Unit)

[@tech_with_moss]

Thanks for your comment and the updates regarding axes!

[@dbopsdev]

Hello, can you please help me here, I can't find Y axis to set the percentile in my Grafana v10.4.2. Where can I find it? this is the query I'm using to capture CPU utilization from Windows exporter: "avg without(cpu) (irate(process_cpu_seconds_total{job="Windows Exporter",instance="Test Server",mode!="idle"}[1m]))"

[@performancetestengineer-go2294]

Great Video really helpfull to me, Thanks . Please do more videos

[@alyonachmil9781]

Thank you. Very useful.

[@mahip2704]

Thanks for info.

[@ocramas]

Not sure if you all are paying attention, but that is a big ass mic right there!

[@rctraveldiaries]

Thanks for the detailed video

[@dbopsdev]

Great video, thank you for sharing the knowledge! Just a question, I can't find Y axis to set the percentile in my Grafana v10.4.2. Where can I find it? this is the query I'm using to capture CPU utilization from Windows exporter: "avg without(cpu) (irate(process_cpu_seconds_total{job="Windows Exporter",instance="Test Server",mode!="idle"}[1m]))"

[@ramithtrevo8241]

Thanks, for the tutorial great explanation , This helped my Assignment a Lot.

[@mzzzd818]

This was crazy helpful. Can you do a video where we look at a potential problem that needs to be solved from monitoring? I am looking for APM work examples. Thanks!

[@tech_with_moss]

Thanks for the kind words and the video suggestion! I definitely will consider doing an APM related video in the future (likely tied to automatic alerting as well)

[@khanetor]

Very useful tutorial! Thanks a lot!

[@adamrozycki9822]

Very helpful, cheers mate!

[@chrisdeez4017]

How do you handle Resource Exhausted? How can I increase the max so that I can avoid any overloading when I query to Grafana? Thanks!

[@noorenshaik6658]

BOOOOOM😂🎉 Too good Man❤

[@taujago]

thank you very much mate. you are legend

[@jobetinfosec3409]

This is by far the best tutorial on Grafana dashboards in YT. Other tutorials are just proposing to install a dashboard already built-up, this one goes into details on how to build it from scratch! BTW when you create a new panel, I noticed that in the right panel, there are 3 default tabs (Panel, Field, Overrides). But looking into other built-up dashboards, I noticed that for the "Gauge" visualization there's only one tab in the right panel ("Panel" tab) with different options than the default ones. I haven't found any documentation or tutorial about this feature... Do you know by chance how to do that?

[@tech_with_moss]

Hello! Thanks a lot. I really appreciate that and I'm glad that you found the tutorial valuable. Regarding your question on Overrides, Panel, and Field, I was able to find some documentation explaining it I think. Field options apply to all fields (columns) in a dataset (such as setting the unit from KB to MB, for instance) and Field Overrides apply to only a single field (column). This is what I understood from the documentation on it: grafana.com/docs/grafana/v7.5/panels/field-options/#field-options grafana.com/docs/grafana/v7.5/panels/field-options/configure-all-fields/#field-option-example grafana.com/docs/grafana/v7.5/panels/field-options/configure-specific-fields/#field-override-example

[@jobetinfosec3409]

@@tech_with_moss Thanks for your feedback! I haven't found yet an explanation for this, but I noticed that I was using the 7.5 version of Grafana, and that the latest version 8.0 has many other features. Maybe that could be related to the version installed... I'll let you know!

[@yoanjouve5673]

Hi ! Thank you a lot, very good explain, great job !

[@kader8815]

can monitor endpoints linux and windows metrics using prometheus and grafana , do you have an idea please ?

[@nubosk]

Hello, thanks for video. I have a question. Why did you use kilobytes/second for network if metric returns bytes?

[@tech_with_moss]

No problem! This was more of a personal preference as I felt like KB was more human-readable than Bytes

[@nubosk]

@@tech_with_moss but it's not about a personal preference. Metric returns bytes, but you say grafana that it's kilobytes. So you will get incorrect Graph.

[@prithivirajan2035]

How do I write a query in Grafana to measure the available space of a linux directory for a specific user id?

[@mahadeoraut1471]

Nice video. I liked the way you present smoothly. I have one dought... Shall I change label Job="anything" Will this work

[@sachinmainkar]

Hello Moss, this is a very very useful video and I am learning all about creating Grafana Dashboards. So THANK YOU so much. One query though. What is that job=Jenkins-job parameter you are using in each panel’s query ?

[@anggasukmabahari6342]

In network traffic query, why you used irate? Why used rate? Thank you for your explanation sir!

[@himel5173]

very helpful....

[@ac3buddy]

Hello, is it possible to monitor process or service state using promql/grafana, for example if sshd is up or down and alert if its down?

[@TyFlowio]

This vid saved my A$$, Thank you!!! :)

[@gopi3825]

Hi moss , is it possible to visualise the kubernetes pod data in grafana , if the data presented (micro service ) in pod is to plot a graph 📉 in that case is it possible to display on grafana

[@priyankaagrawal2611]

Hello Moss, Well explained thanks. I am starting with Grafana, How can I learn about what, when and which metrics query to use? Any recommended documentation?

[@lutfiikbalmajid3128]

Thanks It Works! i guest, better use code instead builder

[@Sanjiocom]

Assuming graphana is querying Prometheus, and prom is querying cpu logs, is there a way to reduce load created by graphana by specifying a max rate of querying, eg poll once every 10 seconds?

[@tech_with_moss]

Great question. Grafana and prometheus should be running on separate infrastructure than the infrastructure that you intend to monitor. The resource data such as CPU and Memory are exposed on the /metrics endpoint by Node Exporter, which is installed on the target infrastructure you want to monitor. You can update Prometheus' configuration for the target endpoint with intervals large enough to not put an unnecessary load on the target infrastructure. This config value in the Prometheus configuration is called the scrape_interval. So prometheus will scrape the /metrics endpoint exposed by Node Exporter according to the scrape_interval you define.

[@Sanjiocom]

@@tech_with_moss Awesome reply. Many thanks indeed.

[@LuisNCalavera]

Muchas gracias

[@coppyphommin9337]

Thank you so much.

[@amandasnasution1987]

what the version of your grafana ?

[@TheSpeechkiss]

Hi Moss, I have typed in node_cpu_seconds_total{job="jenkins-node",instance="localhost:9100", mode!="idle"} into the metrics box but nothing gets loaded. I have a 3 node cluster in aws ( 3 EC2 machines of 2 workers and 1 master) , I have installed Prometheus and Grafana with a helm charts and I can reach both Prometheus and Grafana UI. I know this is not stackoverflow but do you have an idea what could be wrong here ?

[@gouravpattnaik7593]

Hi, This is a really helpful video. Can you please tell how to co- relate multiple queries in a single panel. For example, used and available disk usage in a single panel?

[@mohammadshaazad4131]

Yes, we can do it easily by clicking on the query tab and copy - paste the same code which we are using in another panel.

[@Chatbot121]

Is there a difference between !~ and !=. I saw you use != for one, it may have been a typo

[@Chatbot121]

~16:20 is where you used !=

[@Devops-cloud-school]

Nice content

[@munirawan3609]

Hey I have a question, what if we have 18 physical and 36 logical cores and hyper threading is enabled. In this case, how can we aggregate the cpu utilisation for all cpus/cores?

[@Equality-and-Liberty]

First, i want to thank you for your video. You can explain things very clear. That is better for people to understand. I have a question though. I install Grafana, Prometheus, and node exporter but i don't see the name of Prometheus in my data source field. Do i have to change some settings?

[@tech_with_moss]

Hello! Thank you, I'm glad you liked it! If Prometheus isn't showing up as a data source, I would double check that in the Grafana settings page you have configured the Prometheus server as a new data source and added the Prometheus endpoint in that settings page. I hope this helps.

[@olehharhat3070]

@@tech_with_moss Thanks, i was about to ask same question, all i had to do was setup the Prometheus as a data source in Setting left pannel. Now its working.

[@linuxstudent5158]

Good tutorial, thank you.

[@olehharhat3070]

Thanks, great stuff.

[@MrEric377]

Hello, this is some good content but I had a follow-up question. Would you know if this can also pull process data(On windows)? I a specific list of processes I want it to collect across multiple servers (Sorry, windows servers ). like is that process running???

[@tech_with_moss]

Hi Eric! I'm pretty sure this is possible on windows as well, although I'm not familiar with how to accomplish this. You might want to take a look here: github.com/prometheus-community/windows_exporter . In the README of the Node Exporter GitHub repo (github.com/prometheus/node_exporter ), they recommend this windows exporter utility for windows users

[@sahilbhat4493]

i pulled data using promethues from my linux vm server and allowed it to access from anywhere by opening ports but when i try to login from my windows it gives me this while logging in, "If you're seeing this Grafana has failed to load its application files" , hope youll help me fix it!

[@diegomarado3494]

nice bro, many thanks .

[@userr3953]

Thanks so much for this Moss. I've been trying to figure out how to get my CPU graphs "right side up" since the built-in ones you import (at least for the Windows Node templates) seem to have the graph upside down. I used this exact method to show Windows CPU but had to use "core" instead of "cpu": avg without(core) (irate(wmi_cpu_time_total{instance="10.10.100.103:9182", mode!="idle"}[1m])) , but, it doesn't look like it's accurate, would be nice to get it to look just like the CPU section in Windows task manager -- Performance tab. I don't suppose you could do a tutorial on how to properly show Windows CPU? Thanks again for your videos, I plan to watch them all!

[@krishnaswamydayalan7897]

How to monitor all VM ‘S and getting alert for (cpu usage if crossing threshold 85%)

[@tech_with_moss]

Thanks for your comment! I will likely create a tutorial covering alerts in the future!

[@krishnaswamydayalan7897]

@@tech_with_moss Thanks ..! can you give examples for wmi exporter (Prometheus) ..

[@mohammedadnanmohammed6201]

Good explain ... I have a local quorum network consists of 7nodes ... I use quorum profiling to test the performance of it ... how to get matrics of my network (local quorum network) using grafana?

[@tech_with_moss]

Thank you! Unfortunately I don't really have any familiarity with quorum so I'm not really sure how you could retrieve metrics from that network. Apologies that I couldn't be of more help!

[@jhmc93]

do i need jenkins installed for this

[@akenjifaith3779]

Great Tutorial I created my dashboard by following thee video Please, how can I set up alert rules. For example, alert message when CPU usage above 90%? How do I do this on grafana ?

[@RavGpt]

Thanks Moss. This is wonderful. We were using solarwinds for such tasks however just now we are switching to Grafana, Can we customize it in such a way so that I can create similar dashboard for each VM (~300) we have in our data center? Can we produce such report for each VM we search in Grafana?

[@tech_with_moss]

I'm glad you like it! Yes, it is definitely possible to have a dashboard board like this for N number of virtual machines by utilizing Grafana variables.

[@RavGpt]

@@tech_with_moss Thanks for responding. I tried to search it but couldn't find yet. Can you share some any link that might help me? Thanks again.

[@tech_with_moss]

@@RavGpt I think this video I have on variables might help you: kzbin.info/www/bejne/boDSaYR5a75rpNU

[@RavGpt]

@@tech_with_moss Thank you very much!

[@roberto_camp]

Great job! We are just starting to transition to using Prom/Grafana in our network monitoring suite. Great introduction. Please share if you have any specific dashboards built for switch monitoring with the SNMP-exporter.

[@tech_with_moss]

Thanks I appreciate it! I don't have any switch monitoring dashboards, unfortunately. Good luck with your transition!

[@ichhasseamerika]

Gr8 stuff, thanx!

[@raziffatahillah373]

is this grafana version 7 ?

[@deepaksingh-fx1fx]

i can use grafana for monitoring multiple remote pcs?

[@tech_with_moss]

Yes its definitely possible to monitor multiple remote PCs

[@SteveNgaiCheeWeng]

Hi, I can't find "node_cpu_seconds_total" but I can find "process_cpu_seconds_total". This is due to version?

[@tech_with_moss]

Hi Steve! I'm not entirely sure why node_cpu_seconds_total isn't showing up, but based off this documentation it would seem like both you should be available: github.com/prometheus/node_exporter/blob/master/collector/fixtures/e2e-output.txt . It looks as though the process_cpu_seconds_total only tracks user and system CPU time as oppose to all CPU time modes like IO Wait "# HELP node_cpu_seconds_total Seconds the CPUs spent in each mode. # TYPE node_cpu_seconds_total counter" "# HELP process_cpu_seconds_total Total user and system CPU time spent in seconds. # TYPE process_cpu_seconds_total counter"

[@heyhello-m9t]

thats awesome mate but can u show us how to install node ex[prter and perm from scratch? we are beginerrs not pros!

[@maxfast8868]

good video thanks!

[@harin4]

Hii , its very useful , Can we use Variable in By clause in grafana [sum (node_memory_MemTotal_bytes /( 1024 * 1024 * 1024)) by ( instance )] here i want to use Variable for instance is it possible ??

[@medamineblibech5641]

I Hope to visualisation my Windows 10 Tell me how

[@dedm7759]

красавчик спасибо

[@pratiksawalkar1531]

how to setup node exporter

[@amandubey6898]

Not able to see settings, cause of your videos subtitle

[@walterzz-d1y]

2021 update?

[@tech_with_moss]

I will definitely consider doing an updated dashboard creation video!

[@walterzz-d1y]

@@tech_with_moss great, they upgraded new version of dashboard i cant make the graph

[@tech_with_moss]

@@walterzz-d1y oh OK, that is good to know. Thank you for this info!

[@walterzz-d1y]

@@tech_with_moss wlc

[@singhpandi6849]

This is very helpful thanks. how to find particular PostgreSQL query CPU and memory utilization in grafana dashboards?

[@surajbhanyadav6047]

Did not work for cpu usage

[@MrMurUK]

Downvoting this because I don't need to see you in the corner and you are blocking info. It might not have seemed pertinent at the time but now Grafana is different so it would have been helpful to see the whole screen just to see where things have been moved to.