This repository is meant to provide instant observability to your Kubernetes cluster by only running 2 scripts:
- Installing the New Relic infrastructure agent per Helm
- Deploying New Relic observability per Terraform
- Helm 3.x
- Terraform 0.13+
In order to install New Relic infrastructure agent, you can run the 01_install_infra_agent.sh script with the following arguments:
| Name | Argument Flag |
|---|---|
| deployment name | --deployment |
| deployment namespace name | --deployment-namespace |
| cluster name | --cluster |
--deploymentstands for the name of the Helm deployment (required)--deployment-namespacestands for the namespace name into which the infra agent will be deployed (required)--clusterstands for the name of your cluster with which it will be seen & queried within New Relic (required)
The script will add New Relic charts (https://helm-charts.newrelic.com) to your Helm repositories and will deploy the infrastructure agent
- in priviliged mode
- in low data mode
- with
kube-state-metrics - with
kube-events
Beware
- to define the cluster name unique since this will be considered as one particular cluster by the prebuilt NRQL queries within the Terraform observability scripts
- to define your New Relic license key as an environment variable
NEWRELIC_LICENSE_KEY=xxx
If you want to add more charts within newrelic-bundle, feel free to adapt your installation script as you wish.
After your installation is successful, you will see your Kubernetes data flowing into your New Relic account. In order to have quick & detailed overview to your cluster, you can run the 02_deploy_newrelic_terraform.sh script with the following arguments:
| Name | Argument Flag |
|---|---|
| cluster name | --cluster |
| enrichment | --enrich |
| dry run | --dry-run |
| destroy | --destroy |
--clusterstands for the name of your cluster with which it will be seen & queried within New Relic (required)--enrichstands for wheather you want to enable the workflow enrichment. Defaults to false (optional)--dry-runstands for just runningterraform planand prompting what changes will look like (optional)--destroystands for deleting the Terraform resources (optional)
Beware
- to set the argument
--clusterexactly the same to the one you have defined in your agent installation01_install_infra_agent.sh. - to define your New Relic account ID as an environment variable
NEWRELIC_ACCOUNT_ID=xxx. - to define the region of your New Relic account as an environment
variable
NEWRELIC_REGION="xx".- it is either
usoreu
- it is either
- to define your New Relic API key as an environment variable
NEWRELIC_API_KEY=xxx.
Dashboards contain general information and resource consumption regarding your nodes, namespaces and pods.
You will have 1 dashboard
- for your cluster
- for your namespaces
- for your deployments
- for your daemonsets
- for your statefulsets
This way, you will be able to have many dedicated overviews for every edge of
your cluster. You can inspect every node or namespace individually and navigate
through the dashboards from higher cluster level down to lower pod level. You
can also use of the variables within the dashboards to filter specific
Kubernetes resources (e.g. show only deployment-a and deployment-b within
namespaces namespace-x and namespace-y).
You can refer to the prebuilt queries within the dashboards and deep dive more to get the most out of your telemetry data.
Alerts are currently categorized under nodes and deployments. They are set to fire off whenever a resource utilization (CPU, memory & storage) exceeds warning and critical levels.
In order to be able to monitor the most, it is highly recommended to set requests & limits to your resources. Thereby you can always be aware of the excessive consumption within your cluster.
Workflows are meant to notify you whenever the alert conditions are violated
and cause an issue.
Currently, only emailing is supported.
You can set the local variable emails in the
02_locals.tf
file according to which teams members you want to notify.