spinnaker: ECS deploys broken in release-1.19.x-latest-validated

Issue Summary:

After upgrading to release-1.19.x-latest-unvalidated I was unable to deploy an ECS service

Cloud Provider(s):

ecs

Environment:

local debian

Feature Area:

clouddriver, ecs

Description:

After upgrading to release-1.19.x-latest-unvalidated, ECS deploy (test-v009) was taking forever.

Clouddriver logs show the following log:

Mar 06 22:52:06 ip-172-28-10-142 clouddriver[25314]: 2020-03-06 22:52:06.410  INFO 25314 --- [0.1-7002-exec-7] c.n.s.c.e.p.v.EcsServerClusterProvider   : No ECS Server Groups were found with the name test-v009
Mar 06 22:52:06 ip-172-28-10-142 gate[25305]: 2020-03-06 22:52:06.420  INFO 25305 --- [-serverGroups-9] c.n.s.g.s.internal.ClouddriverService    : <--- HTTP 404 http://localhost:7002/applications/test/serverGroups/stage-ecs/us-east-2/test-v009?includeDetails=false (191ms)

looks as if the newly created servergroup never gets into redis

redis> keys *test-v009*
(empty list or set)
(1.26s)

Steps to Reproduce:

Upgrade to release-1.19.x-latest-unvalidated and start an ECS deploy

Additional Details:

Some weirdness in Deck as well: items in the cluster view disappear periodically