Below are the guidelines for setting up Fabric components and monitoring tools, that support the Monitoring dashboard example, by referring to each of the Fabric workload nodes and to the monitoring machine.
The supportive monitoring tools used for the dashboard example are: Grafana (the example dashboard is being tested on several Grafana versions - 8.3.4 and up), Prometheus, JMX Exporter, Prometheus Node Exporter , Promtail, Loki.
JMX Exporter should be running on all instances. The default configurations should be sufficient.
Node Exporter should be installed and should run on all machines that have the Fabric/Cassandra/Kafka app running.
Grafana should be installed on the monitoring machine.
Prometheus should be installed on the monitoring machine, and it should listen to Node Exporter and JMX.
Log metrics
Promtheus and Loki should be added as data sources in Grafana.
When you start running Promtail, Loki should already be running.
The Dashboard example assumes that the Fabric log files are located on the Fabric cluster nodes at "/opt/apps/k2view/logs/k2fabric.log". In case your deployment is located in another location, change the Fabric Log metric's query. A similar change may be needed in the Promtail config file.
For illustration, below are links to configuration samples for Loki and Promtail. These configurations allow a safe Loki execution, without collecting too much data.
Here you can find a Loki config example.
After making the changes and adjustments, per your deployment, locate it in the Loki installation directory and rename it to "loki-local-config.yaml".
To illustrate a Fabric cluster with 2 nodes, you can find Promtail config example files here for node 1 and here for node 2.
After making the changes and adjustments, per your deployment, locate the Promtail config file, in the Promtail installation directory, and rename it to "promtail-fabric-config.yaml".
The required adjustments are, for example, the machine's IPs:
Below are the guidelines for setting up Fabric components and monitoring tools, that support the Monitoring dashboard example, by referring to each of the Fabric workload nodes and to the monitoring machine.
The supportive monitoring tools used for the dashboard example are: Grafana (the example dashboard is being tested on several Grafana versions - 8.3.4 and up), Prometheus, JMX Exporter, Prometheus Node Exporter , Promtail, Loki.
JMX Exporter should be running on all instances. The default configurations should be sufficient.
Node Exporter should be installed and should run on all machines that have the Fabric/Cassandra/Kafka app running.
Grafana should be installed on the monitoring machine.
Prometheus should be installed on the monitoring machine, and it should listen to Node Exporter and JMX.
Log metrics
Promtheus and Loki should be added as data sources in Grafana.
When you start running Promtail, Loki should already be running.
The Dashboard example assumes that the Fabric log files are located on the Fabric cluster nodes at "/opt/apps/k2view/logs/k2fabric.log". In case your deployment is located in another location, change the Fabric Log metric's query. A similar change may be needed in the Promtail config file.
For illustration, below are links to configuration samples for Loki and Promtail. These configurations allow a safe Loki execution, without collecting too much data.
Here you can find a Loki config example.
After making the changes and adjustments, per your deployment, locate it in the Loki installation directory and rename it to "loki-local-config.yaml".
To illustrate a Fabric cluster with 2 nodes, you can find Promtail config example files here for node 1 and here for node 2.
After making the changes and adjustments, per your deployment, locate the Promtail config file, in the Promtail installation directory, and rename it to "promtail-fabric-config.yaml".
The required adjustments are, for example, the machine's IPs: