Taking out Force-side Redis

[craigspaeth]

Now that Metaphysics is the bee's knees, we should try to kill off all Force-side caching and drop the Open Redis addon.

In most places, it's used to do request-level caching with a cache: true flag when we had spottier performance in Gravity. Now we don't really gain anything with this, and the only places it's still necessary is on pages like the fair microsite that needs to fetch and cache a big blob of data (clear use case for MP), or in often-fetched data like fetching auction reminders which should probably move that performance optimization downstream to Gravity or MP.

Places that need refactoring...

• [ ] Fair microsite (refactor to MP... I hear there's a redesign under works to check in with Katarina & Morgan on)
• [ ] Info menu component (refactor to MP)
• [ ] Venice feature check (kill code + feature or atleast remove caching)
• [ ] EOY list (kill code + feature or atleast remove caching)
• [ ] Auction reminders (remove caching code—move perf. optimization downstream)
• [ ] Partner category fetch (remove caching code—it uses MP)
• [ ] SMS throttling (look at Twillio throttling features)

[anandaroop]

Following along with today's discussion of retiring Open Redis… I'm gonna add to this list a couple of cache: true instances that are for read-only Backbone collections that fetch not from Gravity endpoints, but rather from static files on S3.

• [x] listing of all partner cities
• [x] listing of featured cities

These are used e.g. here and here

Wondering if those also be good candidates to move — do we use Metaphysics to fetch and cache static json blobs?

[craigspaeth]

Thanks @anandaroop ! I think at this point we haven't set any boundaries with MP and any data is fair game in MP, so 👍 to moving that if it feels like the right call to you. However, you could also simply remove the cache: trueflags as those are only mildly increasing performance (caching individual requests) and Gravity + certainly S3 should be performant enough without caching.