Enhancing Kubernetes Controller Documentation for Elastic Agent Deployments

[asmith-elastic]

Proposal

The current documentation does not explicitly guide users on selecting the optimal Kubernetes controller type for deploying Elastic Agents. Users need clear distinctions between DaemonSet, Deployment, and StatefulSet to make informed decisions based on their specific data collection requirements and cluster distribution strategies. Without this guidance, users may not fully capitalize on the capabilities of Elastic Agents within their Kubernetes environments.

Suggestion

To address this gap, it could be helpful to provide an additional section to the existing documenation which highlights the following topis:

When to Use a DaemonSet:

• Scenarios where a DaemonSet is essential for node-specific data collection.
• Advantages of having an Elastic Agent on every Kubernetes node.

When to Use a Deployment:

• Cases where a fixed number of Elastic Agents is optimal.
• The benefits of statelessness and scaling capabilities that Deployments offer.

When to Use a StatefulSet:

• Specific conditions that necessitate stable network IDs or persistent storage.
• Guidance on when the ordered deployment and unique pod identity are critical for data continuity.