Skip to content

Latest commit

 

History

History
101 lines (80 loc) · 3.96 KB

app-readme.md

File metadata and controls

101 lines (80 loc) · 3.96 KB

CaaS Project Monitoring

CaaS Project Monitoring v3 is the evolution of Rancher Project Monitoring V1/V2 and should improve the usage of a monitoring stack with a simple and secure deployment. Based on the community project kube-prometheus-stack the components Prometheus, Alertmanager, and Grafana are installed in the user's namespace.

The deployed prometheus uses the central cluster prometheus' federate endpoint to scrape metrics only related to the user's project. Additional ServiceMonitors, PodMonitors and AlertmanagerConfigs (as well as all other coreos custom resources) can be created in the user's namespace to monitor and alert on the user's resources.

Usage

Installing this chart will install a copy of the kube-prometheus-stack in a rancher project namespace. With default settings at least install namespace are covered with metrics, permitted by the ServiceAccountToken in the Prometheus Pod. If the install namespace is in a project, all other namespaces are also covered by the token permission.

To make sure the deployed prometheus and alertmanager instances can work properly and scrape your ServiceMonitors, AlertmanagerConfigs and all other coreos custom resources as well, make sure to add your projectId to the values.yaml file.

kube-prometheus-stack:
  alertmanager:
    alertmanagerSpec:
      alertmanagerConfigNamespaceSelector:
        matchLabels:
          field.cattle.io/projectId: p-xxxxx
  prometheus:
    prometheusSpec:
      podMonitorNamespaceSelector:
        matchLabels:
          field.cattle.io/projectId: p-xxxxx
      probeNamespaceSelector:
        matchLabels:
          field.cattle.io/projectId: p-xxxxx
      ruleNamespaceSelector:
        matchLabels:
          field.cattle.io/projectId: p-xxxxx
      serviceMonitorNamespaceSelector:
        matchLabels:
          field.cattle.io/projectId: p-xxxxx

To monitor additional Prometheus targets with ServiceMonitors and/or AlertmanagerConfigs in other project namespaces, those namespaces must by listed in caas.projectNamespaces.

Debug

If you need to debug the various components, you can set the log level for each component in the values.yaml file:

kube-prometheus-spec:
  alertmanager:
    alertmanagerSpec:
      logLevel: debug
  grafana:
    grafana.ini:
      log:
        level: debug
  prometheus:
    prometheusSpec:
      logLevel: debug

Preconditions for the cluster operator

  • This app available in app catalog within Git repo or Helm chart repo. Generate an access token with read_repo permissions, if the repo requires authorization.

  • An installed version of CaaS Cluster Monitoring

    • It is important that the coreos CRDs, which are installed in the cluster, are compatible with the kube-prometheus-stack version of this chart. If not, the installation may fail.

  • An installed version of the Navlinks Webhook, to create weblinks of the Project Monitoring in Rancher sidebar menu.

  • The CaaS Project Owner must be assigned to the user installing the app. It provides the additional permissions for the project monitoring stack:

code snippet 
rules:
  - apiGroups:
      - monitoring.coreos.com
    resources:
      - alertmanagers
      - alertmanagerconfigs
      - prometheuses
    verbs:
      - create
      - delete
      - get
      - list
      - patch
      - update
      - view
      - watch

Development

If you need to make any changes to the chart, you should follow this general development process:

  1. Change any files needed for the chart.
  2. Update any dependencies by using the Chart.yaml, if needed.
  3. Update the README.md.gotmpl, if needed.
  4. Update the app-readme.md, if needed.
  5. Run make chart to prepare the chart for packaging.
  6. Open a PR with the changes.