m3db / m3cluster / 1744

86%

push

travis-ci

Add retry options for creating etcd client (#232)

When M3 processes using m3cluster for service discovery comes up, a lot of processes will log a fatal error message and fail if they cannot start up.  

This is fine but can cause a large amount of DNS requests to be issued if restarted by a processes supervisor very frequently.  To avoid some of these thundering herd issues if a KV client is down, we add some level of backoff when trying to connect to service discovery.

21 of 21 new or added lines in 2 files covered. (100.0%)

5964 of 6908 relevant lines covered (86.33%)

38900.07 hits per line