perf: add lru cache

[matthewelwell]

This PR adds LRU caching to the edge proxy which allows users to configure caching on the flags and identities endpoints.

When the edge proxy receives an updated environment document it clears the cache to ensure that updated values are received.

New settings:

{
  ...
  "endpoint_caches": {
    "identities": {
      "use_cache": true | false,
      "cache_max_size": 128
    },
    "flags": {
      "use_cache": true | false,
      "cache_max_size": 128
    }
  }
}

Current concerns: by clearing the cache when we receive an updated environment, we are risking a thundering herd issue in a platform with sustained high load.

An alternative approach (one which can be seen in the first commit of this PR) is to use a ttl_cache instead of one which is completely evicted when an environment update occurs.

I tend to think that the latest approach is better and there are unlikely to be scenarios where the platform is under such significant load that there will be a thundering herd issue, but it's definitely something that should be considered.

In terms of the implementation, it definitely needs some tweaks and some test coverage. I am most concerned about the tight coupling, but I wasn't able to find an easy solution. I will come back to it in the morning...

Finally, I have tested this manually and, for a project that has 300 features, I ran a load test (albeit one that requests the flags for the same identity over and over again) and got an improvement from 137rps to 2600rps for a single edge proxy docker container.

[gagantrivedi]

Approve with one minor comment (and lint fixes)