Visualize Prometheus metrics with Grafana

[s5uishida]

Hi @herlesupreeth

Thank you very much for supporting Prometheus metrics.

I installed Grafana-OSS from DockerHub and was able to view the metrics.

[myonlystarWang]

Hi @s5uishida
Did you follow this guidance completely? (https://open5gs.org/open5gs/docs/tutorial/04-metrics-prometheus/)

[s5uishida]

docker_open5gs contains the necessary settings, so I think you can use it without referring to that tutorial.

[myonlystarWang]

I got it. Thank you very much!

[herlesupreeth]

@s5uishida thank you for trying it out. Integrating with Grafana was on my todo list next :). Will probably ask you if I face any issues integrating prometheus with it.

[herlesupreeth]

@s5uishida btw, you can see graph in prometheus itself. Did you see any benefit in using Graphana vs Prometheus for visualizing graph?

[s5uishida]

Yes (probably).

I think it would be convenient to use it with Grafana to flexibly layout various panels and display them on the dashboard.
I don't know how to flexibly customize dashboards with Prometheus ;-)

[s5uishida]

Will probably ask you if I face any issues integrating prometheus with it.

When access the Prometheus dashboard from the browser of the PC on the network outside DockerHost, you will not be able to see metrics for each job_name (172.22.0.x/24 on Docker).

However, when connect to http://DOCKER_HOST_IP:9090/metrics, you will be able to view all job_name's (NFs) metrics in a browser.

In the case of Grafana, the metrics on the dashboard on the browser will be referenced via the Grafana server, so I think Grafana and Prometheus will communicate within 172.22.0.0/24.

So there is no need to expose ports to the network outside of DockerHost, and the network settings in the current docker-compose.yaml of docker_open5gs are fine.

You can view all NFs metrics from Grafana's Metrics browser (list box). If multiple job_names have the same metric, each job_name is displayed.

For your reference.

[s5uishida]

Hi @herlesupreeth

As a trial, with reference to your docker_open5gs/metrics, I confirmed the operation of visualization of metrics with Grafana as follows.
(as a example)

Access URL: http://DOCKER_HOST_IP:3001/

docker_open5gs/.env

--- .env.orig   2022-11-24 09:48:11.000000000 +0900
+++ .env        2022-11-24 11:45:29.448965227 +0900
@@ -116,3 +116,6 @@
 
 # METRICS
 METRICS_IP=172.22.0.36
+
+# Grafana
+GRAFANA_IP=172.22.0.37

docker_open5gs/docker-compose.yaml

--- docker-compose.yaml.orig    2022-11-24 09:48:11.000000000 +0900
+++ docker-compose.yaml 2022-11-24 11:46:42.013296671 +0900
@@ -658,6 +658,23 @@
     networks:
       default:
         ipv4_address: ${METRICS_IP}
+  grafana:
+    build: ./grafana
+    image: docker_grafana
+    container_name: grafana
+    env_file:
+      - .env
+    volumes:
+      - ./grafana:/mnt/grafana
+      - /etc/timezone:/etc/timezone:ro
+      - /etc/localtime:/etc/localtime:ro
+    expose:
+      - "3000/tcp"
+    ports:
+      - "3001:3000/tcp"
+    networks:
+      default:
+        ipv4_address: ${GRAFANA_IP}
 networks:
   default:
     ipam:

docker_open5gs/grafana/Dockerfile

FROM ubuntu:focal

ENV DEBIAN_FRONTEND=noninteractive

RUN apt-get update && \
    apt-get install -y wget && \
    wget -q https://dl.grafana.com/oss/release/grafana-9.2.6.linux-amd64.tar.gz && \
    tar xvfz grafana-9.2.6.linux-amd64.tar.gz

WORKDIR grafana-9.2.6

CMD /mnt/grafana/grafana_init.sh

docker_open5gs/grafana/grafana_init.sh

#!/bin/bash

bin/grafana-server

[herlesupreeth]

Thank you. Will incorporate shortly. Currently busy with day-to-day job 😅 so will probably work on this over weekend.

[s5uishida]

A little information.

PCF and UPF now support Prometheus. AMF and SMF metrics have also been added a little.
For the list of metrics, please refer to the updated following.

https://github.com/s5uishida/open5gs_5gc_ueransim_metrics_sample_config

[herlesupreeth]

Ah nice!! Thanks for letting me know. Will add them.