A etcd deployment allows for vertical scaling of both the database nodes and its TCP portals.
The vertical scaling option allows for memory and storage to be manually added to each database node. Vertical scaling is the dimension that Compose Auto-scaling operates in.
In the Deployment Resources view for etcd the top panel shows the configured database nodes. Each row displays the node name, its location, and its IP address. In the top right, the memory and storage allocated to the nodes is displayed.
In the lower right of the panel are buttons to adjust the nodes. To add or reduce memory allocated to an etcd deployment, click the Scale button. This takes you to the Scale Nodes view:
By adjusting the slider, you can raise or lower the memory allocated to the etcd nodes. The storage allocation is 1:1 with the selected memory allocation. When the slider is moved, the view updates with the new cost per month for the deployment.
8GB Storage Limit
It is possible to scale your deployment with more than 8GB of memory. However, there is a 8GB limit on storage. If you scale your deployment beyond the 8GB limit you will no longer have persistent storage. Read more about the limit on the Alarms and Status Messages page.
Clicking Confirm triggers the rescaling of the deployment.
For etcd there will be at least two TCP portals. Below is an example from the Scale TCP Portals view where the slider is set to scale them:
Both portals will be scaled to the new memory size.
If you find you do not need the resources currently provisioned to your deployment you can scale down the deployment. Compose auto-scaling will NOT automatically scale down or remove resources that you are not currently using.
Scaling down works similarly to scaling up. Moving the sliders to the left shows the difference in the cost, and the new monthly total.
Use the Security panel to remove any provisioned TCP/HAProxy portals from your deployment.
If this article didn't solve things, summon a human and get some help!
Updated over 3 years ago