Introduction

This chart bootstraps a monitoring system on a Kubernetes cluster using the Helm package manager. This monitor system can be used to monitor an Alluxio cluster started on Kubernetes cluster.

Pre-requisites

Kubernetes

Kubernetes 1.11+ with Beta APIs enabled

Install the Chart

To install the Monitor Chart into your Kubernetes cluster:

$ helm install --namespace "alluxio" "alluxio-monitor" monitor

After installation succeeds, you can get a status of Chart

$ helm status "alluxio-monitor"

Uninstall the Chart

If you want to delete your Chart, use this command:

$ helm delete --purge "alluxio-monitor"

Configuration

The monitor system is implemented based on Prometheus + Grafana, the resource files are placed in the monitor/source directory. Before installing the monitor chart, you may make some appropriate modifications to the configuration.

1. source/grafana/datasource.yaml

This grafana datasource url domain name is [MONITORNAME]-prometheus, for example: our monitor installation name is alluxio-monitor, then it will be 'alluxio-monitor-prometheus'

datasources:
  - name: Prometheus
    ...
    url: http://alluxio-monitor-prometheus:9090

2. source/prometheus/prometheus.yaml

Change each prometheus job's namespace, For example, if the alluxio cluster we want to monitor is installed in alluxio namespace, then edit the prometheus.yaml:

scrape_configs:
  - job_name: 'alluxio master'
    kubernetes_sd_configs:
      - role: pod
        namespaces:
          names:
            - alluxio

3. Enable the alluxio metrics

To use the monitor, we need the alluxio prometheus podAnnotations defined in the '../alluxio/values.yaml' metrics part, so it is necessary to enable metrics before installing the alluxio chart. After that, the monitor can keep track of the target alluxio cluster.

metrics:
  enabled: true
  ...
  PrometheusMetricsServlet:
    enabled: true
  # Pod annotations for Prometheus
  podAnnotations:
     prometheus.io/scrape: "true"
     prometheus.io/port: "19999"
     prometheus.io/jobPort: "20002"
     prometheus.io/workerPort: "30000"
     prometheus.io/path: "/metrics/prometheus/"

4. Download the alluxio dashboard

Download the alluxio dashboard from Alluxio grafana dashboard V1, then move the dashboard file to monitor/source/grafana/dashboard directory.

Helm Chart Values

Full documentation can be found in the comments of the values.yaml file, but a high level overview is provided here.

Common Values:

Parameter	Description	Default
`fullnameOverride`	To replace the generated name	`alluxio-monitor`
`imagePullPolicy`	Docker image pull policy	`IfNotPresent`
`grafanaConfig.name[0]`	Grafana dashboard config name	`grafana-dashboard-config`
`grafanaConfig.path[0]`	Grafana dashboard config path in the image container	`/etc/grafana/provisioning/dashboards`
`grafanaConfig.name[1]`	Grafana datasource config name	`grafana-datasource-config`
`grafanaConfig.path[1]`	Grafana datasource config path in the image container	`/etc/grafana/provisioning/datasources`
`prometheusConfig.name`	Prometheus config name	`prometheus-config`
`prometheusConfig.path`	Prometheus config path in the image container	`/etc/prometheus`

Prometheus values:

Parameter	Description	Default
`imageInfo.image`	The prometheus docker image	`prom/prometheus`
`imageInfo.tag`	The prometheus image tag	`latest`
`port.TCP`	The prometheus default listen address	`9090`
`args`	The prometheus config args, see values.yaml for detail explanation	`--config.file=/etc/prometheus/prometheus.yml --storage.tsdb.path=/prometheus --storage.tsdb.retention=72h --web.listen-address=:9090`
`hostNetwork`	Controls whether the pod may use the node network namespace	`false`
`dnsPolicy`	`dnsPolicy` will be `ClusterFirstWithHostNet` if `hostNetwork: true` and `ClusterFirst` if `hostNetwork: false`	`ClusterFirst`
`resources.limits.cpu`	CPU Limit	`4`
`resources.limits.memory`	Memory Limit	`4G`
`resources.requests.cpu`	CPU Request	`1`
`resources.requests.memory`	Memory Request	`1G`

Grafana values:

Parameter	Description	Default
`imageInfo.image`	The grafana docker image	`grafana/grafana`
`imageInfo.tag`	The grafana image tag	`latest`
`env.GF_AUTH_BASIC_ENABLED`	Environment variable of grafana to enable basic authentication	`true`
`env.GF_AUTH_ANONYMOUS_ENABLED`	Environment variable of grafana to disable anonymous authentication	`false`
`port.web`	The grafana web port	`9090`
`port.hostPort`	The hostPort export node port to visit the grafana web	`8081`
`hostNetwork`	Controls whether the pod may use the node network namespace	`false`
`dnsPolicy`	`dnsPolicy` will be `ClusterFirstWithHostNet` if `hostNetwork: true` and `ClusterFirst` if `hostNetwork: false`	`ClusterFirst`
`resources.limits.cpu`	CPU Limit	`2`
`resources.limits.memory`	Memory Limit	`2G`
`resources.requests.cpu`	CPU Request	`0.5`
`resources.requests.memory`	Memory Request	`1G`