Glad you found it helpful!

Nice, love the simple code-less way! One question though: how to prevent S3 object overwrite?

Yes, using Launch Template is recommended now by AWS, however, at the time of writing, I chose to use LC.

Hey Calvin , the SNS topic is in there because the AutoScaling group will publish a message to it in case of EC2_INSTANCE_TERMINATE. E.g. in case of a scale-down

aaron ueckermann sounds like the S3 bucket is missing the instance-listener.zip I missed this in the article for some reason ☺️ Please copy as mentioned in repo readme #2: https://github.com/stefanolaru/aws-cloudformation-templates/blob/main/typesense-multi-node/README.md

Hi Aaron, did you deploy the network stack before the Typesense cluster stack?

Calvin Butler Awesome, thanks for the detailed reply! 👏

That’s what I had at hand at time of writing, but as you say, doing it with GenerateSecretString can also work and it’s a more elegant solution ;) I would appreciate if you can share how you end up solving it.

That's in UserData to make the service self-healing. In case the instance becomes unhealthy and Autoscaling terminates it and starts a new one, the new instance will be attached the Elastic IP and EBS volume automatically.

Hi Nick, you could add an ALB in front of your instance, though for a single instance and just for SSL certificate it might be overkill. Have a look at my other article about running a multi-node Typesense, it uses a Load Balancer https://blog.stefanolaru.com/how-to-run-highly-available-typesense-cluster-in-aws Another (free) option would be to install Nginx which will handle the SSL termination and proxy to Typesense. But it's very likely possible to manage this directly from Typesense ...