Currently, the counter only shows tasks on READY nodes. It's more correct to also count nodes in the other states except DOWN, because the tasks assocated with those nodes are still assumed to be running for orchestration purposes. One example of when this could matter is during a leader failover when agents are in the process of reconnecting. Signed-off-by: Aaron Lehmann <aaron.lehmann@docker.com> Upstream-commit: 614ad95fbbb7d9f94369a722cb7f557286293fc5 Component: engine
This directory contains code pertaining to the Docker API:
-
Used by the docker client when communicating with the docker daemon
-
Used by third party tools wishing to interface with the docker daemon