Hello, we are planning to start deploy the AWS Opswork soon. how many instances required to start AWS Opswork manage very small scale of nodes (within 10 nodes). Would you please clear us that whether we need 2 EC2 instances or just one EC2 instance that can load AWS Opswork to manage nodes.
Thanks!
I can’t speak for opsworks but a single appropriately sized chef server should be able to handle 1000’s of nodes.
As you are in aws something like an m4.large is generally what I recommend unless you need to scale to 10s of thousands of nodes. You can potentially use something smaller given your deployment size but I would just recommend given the minimal cost just sticking with an m4.large and avoid using t2 instance family types as you may chew up more CPU credits than you might expect.
Useful links:
Are you talking about the newer OpsWorks for Chef Automate (OWCA) or the older OpsWorks Stacks? OWCA let’s you select the server instance type. They recommend a t2.medium can support up to 50 nodes, though I would probably say it’s more like several hundred but it very much depends on your check-in interval since the load isn’t based on node count but converge rate (i.e. 50 nodes converging every 15 minutes == 100 nodes converging every 30 minutes). For OpsWorks Stacks there is nothing to deploy, it’s a fully hosted service run by Amazon.
We mostly will start t2.medium as 50 nodes quota is good for startup. My question is AWS Opswork requires how many AWS EC2 instances to have Chef Automated Stack fully deployed. Since based on Chef Doc, Chef Automate requires Chef Server and Chef Automate Server, so I will take it for grant that we also need 2 instances in AWS to deploy Chef Automate solution. One for Chef server, the other is for Chef Automate server. Please let me know above assumption is correct that AWS Opswork need 2 instance instead of one ( t2.medium). Thanks!
In Automate for Opsworks, the chef server and Automate server are on the same instance, so only one.
Matt,
Thanks!