Update cluster-state-service initial reconciler timeout

[kylbarnes]

When the cluster-state-service starts up, it first runs a reconciler job to sync the ECS state into Etcd. The timeout for this job is currently set to 60s. Depending upon the amount of data stored in Etcd, the Etcd service might take longer than 60 seconds to start, and the cluster-state-service will exit with a Error starting event stream handler: context deadline exceeded. error. We should update the initial reconciler job to allow for a longer timeout, or switch to using an exponential backoff to wait for the Etcd service to fully start before starting the initial reconciler.