Understanding Kubernetes orchestration

Your organization is running containerized applications manually across several virtual machines, but the operations team is spending too much time restarting failed containers, balancing load across servers, and deploying updates without downtime. Which core capability of Kubernetes would most directly address these operational challenges?