Analyze response times of image server for possible optimization
- First, check if response times from various client browser locations dominates over the latency of the call at the server itself. If so, then optimization should revolve around geo-distributing image servers.
- If response time is dominated by internal server response time, compare times of cached/uncached responses and investigate potential changes to caching system.
- Check if internal server response time for avatar fetches is impacted by time to query hivemind for the avatars. If it is, consider possibility of caching the responses (maybe with a local jussi owned by the image server).