When you stop routing requests or pause Elasticsearch cluster nodes, you can perform corrective actions that might otherwise be difficult to complete. For example, if your cluster is being overwhelmed by requests because it is undersized for its workload, its nodes might not respond to efforts to resize. Think of these actions as a maintenance mode for cluster nodes that can stop the cluster from becoming completely unresponsive so that you can resolve operational issues much more effectively.
What these actions accomplish:
- Stop routing: No new requests are routed to the cluster node, but existing requests are permitted to be completed. A gentler, less invasive action than pausing the cluster.
- Pause instance: Suspends the node immediately by stopping the Docker container that the node runs on without completing existing requests. A more direct, aggressive action to regain control of an unresponsive node.
To perform one of these actions on a cluster node:
- Log into the Cloud UI.
On the Deployments page, select your deployment.
Narrow the list by name, ID, or choose from several other filters. To further define the list, use a combination of filters.
- Click Stop routing or click Pause for one of the nodes. Repeat for additional nodes in the cluster, if necessary.
Additionally, you might want to override the RAM to storage disk capacity for a deployment if your deployment is unhealthy or at capacity. By opening the reset quota window and changing the ratio, you can update all of the nodes in the deployment that have the same instance type. If the cluster was being overwhelmed by requests, a possible next step might be to resize the deployment, after which the new cluster nodes will start accepting requests again automatically. For other corrective actions, you might need to click Start routing or click Resume after your changes are complete.