Installation

Install Datafy

Install Datafy on any nodes or instances with attached EBS volumes that you want Datafy to monitor or manage.

All installations include two parameters:

• $VERSION or $HELM_VERSION - The version number of the agent (in EC2) or Helm chart (in K8s) to be installed. We recommend always specifying it, for consistent behavior.

• $TOKEN - Security token used for identification, associated with the relevant AWS account. The security token can be generated by an admin through the Datafy app, or omitted entirely with tokenless installation.

If you are installing on an instance where Datafy is already installed, check out how to upgrade

Installing behind an HTTPS proxy? See Network Access.

EC2 Instance

Datafy is installed on EC2s as a Docker image. If not already installed on your instance, install Docker before installing Datafy.

Run the following CLI command on your instance to install Datafy:

Sensor

curl -sSfL https://agent.datafy.io/install?version=$VERSION | AGENT_MODE="sensor" TOKEN=$TOKEN sh

AutoScaler

curl -sSfL https://agent.datafy.io/install?version=$VERSION | AGENT_MODE="autoscaler" TOKEN=$TOKEN sh

The install command downloads and runs an install script that ensures your instance is supported by Datafy, and then installs the requested version using Docker.

Check out Deployment Examples for examples how to install Datafy with the deployment tools you already use

Kubernetes Cluster

You can install Datafy using a single line command or with a Helm chart directly.

When installing AutoScaler, make sure the nodes you install on are supported. If your cluster has a mix of supported and unsupported nodes, see Kubernetes Node Scheduling.

Helm Chart Installation

Datafy can be installed directly via a Helm chart. See the latest Helm chart documentation for additional installation parameters and versioning.

The Helm chart installs the Datafy agent as a DaemonSet, by default on every node in your cluster (see Kubernetes Node Scheduling to install on a subset of nodes instead). For AutoScaler installations, it also installs a Datafy controller Deployment, and adds a datafy-proxy sidecar container to your existing EBS CSI driver pods. See AutoScaler on Kubernetes for what each component does. If the installation fails, Helm will roll back the cluster to the state it was in before the installation began.

For AutoScaler installations, we recommend adding a node taint to gate scheduling around node startup and AutoScaler outages. See Node Taints.

1

Add Datafy Repo

helm repo add datafyio https://helm.datafy.io/datafy-agent
helm repo update

2

Install Datafy

Sensor

helm upgrade --install datafy-agent \
--version "$HELM_VERSION" datafyio/datafy-agent \
--namespace datafy-agent --create-namespace \
--set agent.mode="sensor" \
--set agent.token="$TOKEN" \
--atomic   

AutoScaler

helm upgrade --install datafy-agent \
--version "$HELM_VERSION" datafyio/datafy-agent \
--namespace datafy-agent --create-namespace \
--set agent.mode="autoscaler" \
--set agent.token="$TOKEN" \
--atomic

To successfully install Autoscaler on nodes running Bottlerocket, you must disable kernel lockdown prior to installation. Add the following configuration to the node's user data:

[settings.kernel]
lockdown = "none"

To read the token from a Secret you already manage, replace agent.token with agent.externalTokenSecret.name and agent.externalTokenSecret.key. Both are required.

Single-Line Installation

If your cluster is running Bottlerocket, use the Helm chart installation above

Sensor

curl -sSfL https://agent.datafy.io/install-k8s?version="$HELM_VERSION" | AGENT_MODE="sensor" TOKEN="$TOKEN" sh

AutoScaler

curl -sSfL https://agent.datafy.io/install-k8s?version="$HELM_VERSION" | AGENT_MODE="autoscaler" TOKEN="$TOKEN" sh

Validate Installation

You're all set up!

Once you've successfully completed the installation, you can see the utilization of the volumes attached to instances Datafy is installed on in the Datafy app.

If you've installed Autoscaler, you can now use the "AutoScale" toggle or create an autoscaling rule to activate autoscaling for the selected volumes.

To try that on a test volume first, follow My First Autoscaling Volume.