Increase database server instance sizes

[GUI]

As we discussed today, we'll be increasing the database instance sizes as a temporary measure to help with analytics performance while we work on a better long-term solution (#235).

Currently we're on m4.xlarges, so moving to m4.2xlarges should give us a pretty good boost. We'll also need to adjust the ElasticSearch config to allocate more memory towards it.

This swap shouldn't result in any downtime for API traffic, but it will result in a couple of brief outages of the analytics queries in the admin interface. I'll plan to do this upgrade later this evening or over the weekend to minimize impact for any admin users.

[GUI]

The servers have been upgraded.

I'm running some optimize commands on some of the static data from last year, so that's chugging along in the background. The optimizations seem to be taking longer than I expected, so those might still be running for a couple days. But even with those running in the background, the performance of the analytics in the admin is better, so I think we should be good for a while.