Simple monitoring for workadventure and jitsi.
It consists of prometheus, grafana and caddy. Caddy is used due to it simple revers proxy function and automatic LE cert retrieval and renewal.
- workadvanture installtion
- jitsi server
- https://github.com/prometheus/node_exporter on the WA host and the jitsi host
WARNING: All stats that are polled are unencrypted and unauthenticated. Make sure to switch to TLS and basic auth. Securing your stats is not part of this document.
First you need to install the prometheus-jitsi-meet-exporter on your jitsi server. Please follow the instructions to activate the colibri stats in jitsi.
If you run the jvb as a docker container you could add these to the JVB
environment:
JVB_ENABLE_STATISTICS=true
JVB_STATISTICS_TRANSPORT=muc,colibri
And also add this to the docker-compose.yml
promexport:
image: systemli/prometheus-jitsi-meet-exporter:latest
restart: ${RESTART_POLICY}
command: -videobridge-url http://jvb:8080/colibri/stats
ports:
- '9888:9888'
depends_on:
- jvb
networks:
meet.jitsi:
After the prometheus exporter has been started you should be able to get the statistics via http://localhost:9888/metrics
.
git clone https://github.com/spezi/workadventure_monitor.git
cd workadventure_monitor
./pre-run.sh
Now you need to modify your prometheus/etc/prometheus.yml and set the correct domain names.
Make sure to change the grafana.example.com
hostname to your machines hostname in docker-compose.yml
Once you are done you should be able to start it via
docker-compose up -d
After a short while you should be able to login into grafana with admin
admin
.
Then you add the prometheus datasource with URL prometheus:9090
.
Once this is done you just add this dashborad
And if everything works you can just create another "graph" panel and use workadventure_nb_clients_per_room
as PromQL query.
You can check out the full license here
This project is licensed under the terms of the MIT license.