monitoring

Monitoring Namespace

This Namespace adds software for log and metrics collection and visualization, as well as alarmistic.

Releases

• kube-prometheus-stack
  Installs the kube-prometheus stack, a collection of Kubernetes manifests, Grafana dashboards, and Prometheus rules
• loki
  Helm chart for Grafana Loki in microservices mode
• node-problem-detector
  This chart installs a node-problem-detector daemonset. This tool aims to make various node problems visible to the upstream layers in cluster management stack.
• promtail
  Promtail is an agent which ships the contents of local logs to a Loki instance

Features

• Actively maintained by GraphOps and contributors
• Common values interfaces across all namespaces
• Flexible and adaptable, allowing defaults to be overridden
• Two release channels: stable and canary
• A large selection of Namespaces (listed below)

Quickstart

Note Launchpad Starter is a great way to make use of Namespaces and worth checking out as a starting point for every new Launchpad deployment.

To use Namespaces you will require both a Kubernetes cluster and Helmfile. As such:

• Make sure your Kubernetes Cluster is in order and your environment has the kubeconfig context adequately setup
• Install helmfile, upstream guidance available here: Helmfile Installation – Install kustomize, upstream guidance available here: Kustomize Installation. Although launchpad–namespaces doesn't explicitly use kustomize, it is a dependency for utilising helmfile features.

Next, setup an helmfile.yaml file that makes use of the Monitoring Namespace by creating it with the following contents:

helmfiles:
  - path: git::https://github.com/graphops/launchpad-namespaces.git@monitoring/helmfile.yaml?ref=monitoring-latest
    selectorsInherited: true

Note On the path to the helmfile, you can use the query string's ref (?ref=monitoring-latest) to track one of the release streams: stable and canary, pin to a specific version or just track a particular major or minor semantic version. For more on this, check the Updates section

This is a very minimalist helmfile but enough to get it done. Proceed by running helmfile:

helmfile sync -i

After some output, you should be greeted by a prompt like this:

Do you really want to sync? Helmfile will sync all your releases, as shown above.

[y/n]:

Answer 'y' and hopefully the installation will conclude successfully.

overriding namespace and releases' values

To customize the configuration and deployment, you can pass values to override the default helmfile configuration like so:

helmfiles:
  - path: git::https://github.com/graphops/launchpad-namespaces.git@monitoring/helmfile.yaml?ref=monitoring-latest
    selectorsInherited: true
    values:
      targetNamespace: "i-choose-my-own-namespace"
      labels:
        awesome.label.key/stuff: "yes"
        awesome.label.key/thing: "kind-of-thing"

where we add some labels to this Namespace releases, and set it to be deployed on cluster namespace different from default.

You can also easily override values for every release, like so:

helmfiles:
  - path: git::https://github.com/graphops/launchpad-namespaces.git@monitoring/helmfile.yaml?ref=monitoring-latest
    selectorsInherited: true
    values:
      targetNamespace: "i-choose-my-own-namespace"
      labels:
        awesome.label.key/stuff: "yes"
        awesome.label.key/thing: "kind-of-thing"
      <release-name>:
        - akey: value
          bkey: value

Check out the Namespaces list below for release names, and each chart's folder for its specific values interface.

To use multiple namespaces on the same cluster, just add more items to the helmfiles array like so:

helmfiles:
  - path: git::https://github.com/graphops/launchpad-namespaces.git@monitoring/helmfile.yaml?ref=monitoring-latest
    selectorsInherited: true
    values:
      <monitoring values>
  - path: git::https://github.com/graphops/launchpad-namespaces.git@<other namespace>/helmfile.yaml?ref=<other namespace>-latest
    selectorsInherited: true
    values:
      <other values>

Values

Key	Type	Default	Description
annotations	object		Add annotations to release resources on this namespace
features	list of strings	[metrics, logs]	enum of:  (metrics | logs)
kube‑prometheus‑stack	object
kube‑prometheus‑stack.annotations	object		Add annotations to resources on this release
kube‑prometheus‑stack.chartUrl	string		Override this release's chart URL (i.e: an absolute like /path/to/chart.tgz or /path/to/chart_dir. Or a remote like git::https://github.com/bitnami/charts.git@bitnami/apache?ref=main)
kube‑prometheus‑stack.chartVersion	string		Specify a specific chart version to use for this release
kube‑prometheus‑stack.labels	object		Adds helmfile labels to this release
kube‑prometheus‑stack.mergeValues	boolean	true	Merges passed values with namespace's defaults if true, overrides if false
kube‑prometheus‑stack.resourceLabels	object		Adds labels to resources on this release
kube‑prometheus‑stack.values	(object or list of objects)		Pass values to the release helm chart
kubeVersion	string		Specifies the kubernetes API version, useful in helm templating environment
labels	object		Adds helmfile labels to releases on this namespace
loki	object
loki.annotations	object		Add annotations to resources on this release
loki.chartUrl	string		Override this release's chart URL (i.e: an absolute like /path/to/chart.tgz or /path/to/chart_dir. Or a remote like git::https://github.com/bitnami/charts.git@bitnami/apache?ref=main)
loki.chartVersion	string		Specify a specific chart version to use for this release
loki.labels	object		Adds helmfile labels to this release
loki.mergeValues	boolean	true	Merges passed values with namespace's defaults if true, overrides if false
loki.resourceLabels	object		Adds labels to resources on this release
loki.values	(object or list of objects)		Pass values to the release helm chart
node‑problem‑detector	object
node‑problem‑detector.annotations	object		Add annotations to resources on this release
node‑problem‑detector.chartUrl	string		Override this release's chart URL (i.e: an absolute like /path/to/chart.tgz or /path/to/chart_dir. Or a remote like git::https://github.com/bitnami/charts.git@bitnami/apache?ref=main)
node‑problem‑detector.chartVersion	string		Specify a specific chart version to use for this release
node‑problem‑detector.labels	object		Adds helmfile labels to this release
node‑problem‑detector.mergeValues	boolean	true	Merges passed values with namespace's defaults if true, overrides if false
node‑problem‑detector.resourceLabels	object		Adds labels to resources on this release
node‑problem‑detector.values	(object or list of objects)		Pass values to the release helm chart
promtail	object
promtail.annotations	object		Add annotations to resources on this release
promtail.chartUrl	string		Override this release's chart URL (i.e: an absolute like /path/to/chart.tgz or /path/to/chart_dir. Or a remote like git::https://github.com/bitnami/charts.git@bitnami/apache?ref=main)
promtail.chartVersion	string		Specify a specific chart version to use for this release
promtail.labels	object		Adds helmfile labels to this release
promtail.mergeValues	boolean	true	Merges passed values with namespace's defaults if true, overrides if false
promtail.resourceLabels	object		Adds labels to resources on this release
promtail.values	(object or list of objects)		Pass values to the release helm chart
resourceLabels	object		Adds labels to release resources on this namespace
targetNamespace	string	monitoring	Sets the cluster namespace in which the releases will be deployed
helmDefaults	object
helmDefaults.args	list of strings
helmDefaults.cleanupOnFail	boolean
helmDefaults.createNamespace	boolean
helmDefaults.force	boolean
helmDefaults.historyMax	number	10	limit the maximum number of revisions saved per release. Use 0 for no limit.
helmDefaults.kubeContext	string
helmDefaults.recreatePods	boolean
helmDefaults.tillerNamespace	string
helmDefaults.tillerless	boolean
helmDefaults.timeout	number
helmDefaults.tls	boolean
helmDefaults.tlsCACert	string
helmDefaults.tlsCert	string
helmDefaults.tlsKey	string
helmDefaults.verify	boolean
helmDefaults.wait	boolean
helmDefaults.waitForJobs	boolean