Telescope: 0.9

We're getting close to 1.0 release, and it's time to add important features so we can get a robust, versatile and also easy to use Telescope 1.0.

For this release I worked on some interesting issues:

Unreachable feed list
Telescope gets a list of feeds from a server that sometimes it's not available, and when that happens it causes Telescope to crash. I added a small fix and now if the server is not available, Telescope will keep trying to get the list from the server with little delays in between attempts.

Nginx
I also worked with @c3ho to make some changes to our reverse proxy, and now Telescope uses nginx to serve static content and cache its endpoints. The interesting part is that we had to figure out the right settings to cache only certain endpoints to avoid security issues. Check out some other interesting changes he made to our nginx settings.

Persistent Storage
I liked this one. It was a tiny change, but before this addition Telescope had to process all our feeds (and posts) every time a new PR was merged to master. Now we have persistent storage, so our processed content gets stored on disk and fresh Telescope deployments can use that stored content to show fully updated feeds and posts after every merge.

For 1.0
I also wanted to add something similar to what I did for Elasticsearch, but for Redis. Currently, Telescope waits for Elasticsearch to be ready before start processing feeds, but it doesn't wait for Redis. This could cause data loss or even unexpected crashes, so I'll be working on this for 1.0.