The code he pasted at 8:44 : sum by(instance) (rate(demo_cpu_usage_seconds_total{mode!="idle"}[5m])) / on(instance) group_left() demo_num_cpus 10:10 topk(3, sum by(path, method) (rate(demo_api_request_duration_seconds_count[5m])))
Julius, as a long-time explainer and trainer on OSS, you're doing some seriously great work here, simply explaining things so others can have a good experience and follow along until they can fly themselves, thank you!
I'm using Arch Linux with xfce4-terminal and zsh as a shell. The zsh prompt theme is the popular "agnoster" theme, which creates the arrows in the prompt (and also gives visual git status indications and so on). Needs a powerline font to display correctly. Window manager is i3.
Hey Julius, quick question, I'm currently pursuing the understanding-promql training in promlabs, but I'm especially curious about PromLens as an interface. I can't find any tutorial, just a live demo and some querying guidance from the understanding-promql training. Its there any material, or any sort of walkthrough the interface more than the landing page? I would like to understand how to take the most of promlens. Thank you!
Hey hii Can you just let me know if we can create an alert for jobs.. If the process doesn't complete on time can we create an alert which will notify the developer
Hello, Thank you for the tutorial. It was very informative. I have one question : I want to add a column in the table which shows names of the nodes (for my blockchain network). The dashboard template I am using is Besu Overview by Grafana. How do I do it ?
Wow, Julius, this was so helpful thank you! One question though, can you set up alerts in Grafana to notify you when certain thresholds are breached for any of the metrics being monitored?
Hi Beatrice, glad you liked it! Yes, that's possible in Grafana, take a look at grafana.com/docs/grafana/latest/alerting/ and grafana.com/docs/grafana/latest/alerting/fundamentals/. You can either configure "Grafana-managed rules" that can alert in a generic way on data from any data source, or "Mimir and Loki rules" that work with Mimir (a Prometheus-compatible data source) and Loki (for logs). Maybe I'll make a future video about that! Personally, I still like to configure native Prometheus alerting rules directly in Prometheus itself.
Hello Julius, is it possible to delete one gauge out of the 3? we have a requirement that one gauge is shows the archived logs. we need the real time log to be displayed in gauge.
Hi Manoj, it's a bit hard to tell from the level of detail in your question, but it sounds like maybe you want to expose gauges not about something that is happening inside your own program, but about log files that exist on a system somewhere? In that case, it sounds like you might want to build an exporter rather than directly instrumenting a software itself, see: prometheus.io/docs/instrumenting/writing_exporters/#writing-exporters. I also go into exporters in this training module: training.promlabs.com/training/understanding-and-building-exporters. For an exporter, you basically re-create all exported metrics from scratch every time the exporter is scraped, so each time you will only output exactly the time series that are relevant for that particular scrape and not store any state in between scrapes. If that's not quite what you're looking for, I can also recommend sending a more detailed question to the users mailing list or one of the other open community channels (see prometheus.io/community/), and I or others would be happy to help there as well.
The code he pasted at 8:44 :
sum by(instance) (rate(demo_cpu_usage_seconds_total{mode!="idle"}[5m]))
/
on(instance) group_left() demo_num_cpus
10:10
topk(3, sum by(path, method) (rate(demo_api_request_duration_seconds_count[5m])))
Julius, as a long-time explainer and trainer on OSS, you're doing some seriously great work here, simply explaining things so others can have a good experience and follow along until they can fly themselves, thank you!
Awesome, huge fan of prometheus (still a beginner) , the simple architecture and the kind of services it can provide, too good
Thankyou for such an awesome and crisp explanation. Please make more videos areound promql and grafana
very very helpful, ty very much Sir. I will take a time in your channel to watch more staff like this :)
What distro/terminal application are you running here? I've not seen one like it before, I love how it uses arrows and stuff for the paths
I'm using Arch Linux with xfce4-terminal and zsh as a shell. The zsh prompt theme is the popular "agnoster" theme, which creates the arrows in the prompt (and also gives visual git status indications and so on). Needs a powerline font to display correctly. Window manager is i3.
it's really helpful for me.👏
i was struggling earlier now i have a decent idea of those dash things, thanks man
i can't believe this all stuff is open source, Thanks very much!
Hey Julius, quick question, I'm currently pursuing the understanding-promql training in promlabs, but I'm especially curious about PromLens as an interface. I can't find any tutorial, just a live demo and some querying guidance from the understanding-promql training. Its there any material, or any sort of walkthrough the interface more than the landing page? I would like to understand how to take the most of promlens. Thank you!
Saw your email and replied there :)
Thanks for writing and open-sourcing such great software, Julius.
Thank you. This video helped me understand how to make table as I need.
Wonderfull vid! thanks for the tutorial
Thank you so much for this video, this is very helpful
The gauge one helped me a lot :D thank you.
Simple and concise. Great video once again.
Great job, very helpful!
amazing video!!! please continue!!!
Hey hii
Can you just let me know if we can create an alert for jobs..
If the process doesn't complete on time can we create an alert which will notify the developer
Hello, Thank you for the tutorial. It was very informative. I have one question : I want to add a column in the table which shows names of the nodes (for my blockchain network). The dashboard template I am using is Besu Overview by Grafana. How do I do it ?
From VietNam. Thanks
Thank you so much .. this very informative
Please can you make a video .. how we can copy a panel from existing dashboard. to another dashboard
Wow, Julius, this was so helpful thank you!
One question though, can you set up alerts in Grafana to notify you when certain thresholds are breached for any of the metrics being monitored?
Hi Beatrice, glad you liked it!
Yes, that's possible in Grafana, take a look at grafana.com/docs/grafana/latest/alerting/ and grafana.com/docs/grafana/latest/alerting/fundamentals/. You can either configure "Grafana-managed rules" that can alert in a generic way on data from any data source, or "Mimir and Loki rules" that work with Mimir (a Prometheus-compatible data source) and Loki (for logs). Maybe I'll make a future video about that! Personally, I still like to configure native Prometheus alerting rules directly in Prometheus itself.
How to add DB connections dashboard?
Thank you, Julius, for sharing such a great tutorial and well explanation of how to create and modify graphs.
How do you not blink in the into?
Fantastic video. Thank you!
i did everything same but metrics does not match metrics in system monitor.
very very helpful ty ty
sir
Hello Julius, is it possible to delete one gauge out of the 3? we have a requirement that one gauge is shows the archived logs. we need the real time log to be displayed in gauge.
Hi Manoj, it's a bit hard to tell from the level of detail in your question, but it sounds like maybe you want to expose gauges not about something that is happening inside your own program, but about log files that exist on a system somewhere? In that case, it sounds like you might want to build an exporter rather than directly instrumenting a software itself, see: prometheus.io/docs/instrumenting/writing_exporters/#writing-exporters. I also go into exporters in this training module: training.promlabs.com/training/understanding-and-building-exporters. For an exporter, you basically re-create all exported metrics from scratch every time the exporter is scraped, so each time you will only output exactly the time series that are relevant for that particular scrape and not store any state in between scrapes. If that's not quite what you're looking for, I can also recommend sending a more detailed question to the users mailing list or one of the other open community channels (see prometheus.io/community/), and I or others would be happy to help there as well.
thank you, really useful.
This was very informative and easy to follow. Thank you!