Send metrics from existing or new Prometheus installations with remote_write
Steps for configuring Prometheus and
remote_write to ship metrics to Grafana Cloud depend on how you installed Prometheus into your Kubernetes clusters.
Note: Instead of manual configuration, you can use Kubernetes Monitoring to guide you through the configuration process, allowing you to scrape kube-state-metrics, collect Kubernetes events, and use preconfigured dashboards and alerts. See Configure with Agent Flow mode for more information.
The following are options if you have an existing Prometheus installation.
Configure for Prometheus Operator or kube-prometheus
Configure for Helm and kube-prometheus-stack
When Prometheus Operator and kube-prometheus are deployed using the Helm kube-prometheus-stack chart, follow the steps to Configure remote_write with Helm and kube-prometheus-stack.
Configure for Helm and Prometheus
Configure for Prometheus ConfigMap
When Prometheus is deployed using a Kubernetes Deployment and ConfigMap manifests, follow the steps to Configure remote_write with a Prometheus ConfigMap.
New Prometheus installations
If you are installing Prometheus Operator for the first time into a Kubernetes cluster to send metrics to Grafana Cloud, use one of the following methods.
Migrate a kube-prometheus Helm stack
You can install the kube-prometheus-stack Helm chart into your cluster, and send metrics to Grafana Cloud. You can also use the allowlist to reduce metrics usage, and migrate dashboards and rules to Grafana Cloud. For more details, see Migrate a Kube-Prometheus Helm stack to Grafana Cloud.
Install Prometheus Operator with Grafana Cloud for Kubernetes
You can install Prometheus Operator in a Kubernetes cluster, configure it to scrape an endpoint, and send scraped metrics to Grafana Cloud, follow the steps in Install Prometheus Operator with Grafana Cloud for Kubernetes.