April 2018
Intermediate to advanced
468 pages
14h 34m
English
Kubernetes will take non-responsive pods out of service within a few seconds. The federation control plane monitors the health of clusters and the endpoints behind all of the shards of your federated service in the different clusters. It will take them in and out of service as required, for example, when all of the endpoints behind a service, an entire cluster, or a whole availability zone go down. The inherent latency of DNS caching (3 minutes for Federated Service DNS records by default), may send failover of clients to an alternative cluster in the case of catastrophic failure. However, given the number of discrete IP addresses that can be returned for each regional service endpoint ...
Read now
Unlock full access