Description

This program backs up all grafana dashboards from a given server to S3. The user provides the Grafana server URL and the destination S3 Bucket.

Setup

To build the executable, use make build. To build a docker image from this executable, use make image. Note: make image does not automatically build the executable.

Usage

The easiest way to test this program (on a computer with AWS CLI access configured in ~/.aws) is to run the command:

Basic Testing

./grafana-backup-runner -grafanaURL ${GRAFANA_URL} -s3Bucket ${S3_BUCKET_NAME} -useSharedConfig

Running on Kubernetes

To use this program on a production Kubernetes cluster, the user must define an appropriate set of secrets with the relevant AWS credentials. It is recommended to place the secrets in an env file like so:

AWS_ACCESS_KEY_ID=SUPER_SECRET
AWS_SECRET_ACCESS_KEY=SUPER_SECRET
AWS_REGION=us-east-1

GRAFANA_BACKUP_GRAFANA_URL=YOUR_BACKUP_URL
GRAFANA_BACKUP_S3_BUCKET_NAME=YOUR_BUCKET_NAME

This env file can be used to create the required Kubernetes secret via the command:

kubectl create secret generic grafana-backup-runner-config --from-env-file grafana-backup-runner-config.env

To run a single backup operation on a Kubernetes cluster, use the following manifest:

kubectl create -f manifests/grafana-backup-runner-job.yaml

Once this job completes successfully, you should see a directory with a name of the form "grafana_backup-$TIMESTAMP" appear in your S3 bucket.

Recurring Backups

In order to use Kubernetes' CronJob support to implement recurring backups, the CronJob alpha feature must be enabled on the Kubernetes API server. To test if your API server supports CronJob, run kubectl api-versions. If you do not see batch/v2alpha1 in the output, you will need to enable it yourself. For a cluster deployed using Kubespray, this can be accomplished by editing the appropriate manifest file:

sudo vi /etc/kubernetes/manifests/kube-apiserver.manifest

Add the following line to the apiserver invocation:

- --runtime-config=batch/v2alpha1

You may need to restart the API server for this setting take effect.

Once the CronJob-compatible API version is successfully enabled, you can deploy a recurring backup CronJob using our provided manifest:

kubectl create -f manifests/grafana-backup-runner-cronjob.yaml