Test

Fix two separate but related problems with watch retry handling #820

Job	Run time
Benchmark	2m 25s
Lint	8s
Main (7.2, 3.3)	1m 0s
IPS	2m 29s
Main (7.2, 3.3, compose.ssl.yaml)	1m 12s
Massive Cluster	2m 49s
NAT-ted Environments	55s
Main (7.2, 3.3, hiredis)	49s
Profiling (single)	42s
Main (7.2, 3.3, hiredis, compose.ssl.yaml)	1m 10s
Profiling (excessive_pipelining)	41s
Main (7.2, 3.3, 2, compose.replica.yaml)	1m 1s
Profiling (pipelining_in_moderation)	36s
Main (7.2, 3.2, compose.auth.yaml)	1m 0s
Main (7.0, 3.1)	55s
Main (6.2, 3.0)	55s
Main (5.0, 2.7)	1m 4s
Main (jruby)	1m 31s
Main (truffleruby)	1m 26s
Main (test_cluster_state, 7.2, 2, compose.replica.yaml, 9)	6m 18s
Main (test_cluster_state, 6.2, 2, compose.replica.yaml, 9)	6m 1s
Main (test_cluster_broken, 7.2, no, 6)	1m 54s
Main (test_cluster_broken, 6.2, no, 6)	2m 20s
Main (test_cluster_scale, 7.2, compose.scale.yaml, 8)	1m 52s
Main (test_cluster_scale, 6.2, compose.scale.yaml, 8)	1m 43s
	42m 56s

Provide feedback