Set up monitoring

We should set up monitoring so we know when www_alum goes down, and if we end up with any major deliverability problems. Here are two ways that I've thought of that we could do this:

- Configure www_alum and Caddy to log to syslog (postfix already does this), and then forward our syslog to Papertrail. From there, we can set up email alerts for some error messages.
- Set up Pingdom or something like it to see if https://alum.recurse.com is reachable from the outside.

Something else?