Stop using AWS's elasticsearch service

### Context

AWS Elasticsearch somehow takes forever to spin up. I had a fresh deploy fail because ES didn't come up within an HOUR:

```
aws_elasticsearch_domain.es: Still creating... (1h0m10s elapsed)

Error: Error applying plan:

1 error(s) occurred:

* aws_elasticsearch_domain.es: 1 error(s) occurred:

* aws_elasticsearch_domain.es: "arn:aws:es:us-east-1:589864003899:domain/es-kurt-dev": Timeout while waiting for the domain to be created
```

### Problem or idea

I'm pretty sure ES is just a server with some software installed on it. I cannot imagine what's taking so incredibly long, but I bet we can do it faster if we just install it ourselves. Also according to https://code972.com/blog/2017/12/111-why-you-shouldnt-use-aws-elasticsearch-service there's a lot of other issues with the service and it's twice as expensive.

### Solution or next step

Use an EC2 instance to host Elasticsearch for us rather than AWS's provided service.

Start by locating the infrastructure configuration that defines aws_elasticsearch_domain.es and the deployment path that waits for it; the issue provides no file or test path. Done means managed AWS Elasticsearch is replaced by an EC2-hosted Elasticsearch setup and a fresh deploy completes without the managed-domain timeout.

Written by the indexing model from the issue text.