Track Redis RTT and log if poor

[mperham]

Redis round trip time or network latency can make a big difference in Sidekiq performance. A quick informal Twitter survey shows that a good, healthy Redis service in AWS will exhibit RTT < 1000µs. A cheapie, free service shows latency between 5-20ms. Warn if RTT goes over 50ms.

For now I don't have a good plan to expose this data in the Web UI but it is available in the API:

Sidekiq::ProcessSet.new.each { |sp| puts sp["rtt_us"] }

          Your Redis network connection is performing extremely poorly.
          Current RTT is #{rtt}µs, ideally this should be < 1000.
          Ensure Redis is running in the same AZ or datacenter as Sidekiq.