We are building an app in AWS, using Node.js, for which we want to do continous deployment. We are using DynamoDB for the backend. We would like to use the URL swap for 0 downtime conversions.
I'm trying to figure out the best way to do this. I can create a configuration using the node.js sdk that I can use to make a second environment, or use the eb CLI tool make a clone. But I don't know what resources these cover - for instance, if I specify a multi-server environment from 2 to 4 servers, it'll start with 2. If the load increases to the point where we have 3 servers running, and I use clone or create a configuration, I believe that the new environment is only going to have two servers. Also, not sure about changes we make to alarms and security groups.
Anyone done this, and found a good way to do it? I'm pretty sure I can just slog through each type of resource, retrieving what's already there and adding them to the create scripts. But that's fairly complex.
We have considered using the rolling update of what's in place, reusing the environment that is there. This is relatively easy, as it leaves everything as is, just updates the code. But then we have to manage what is going on with the data as some folks come in with the new code, and some come in with the old.
Aucun commentaire:
Enregistrer un commentaire