Hands-On Microservices with Kubernetes

Self-healing with Kubernetes

Self-healing is a very important property of large-scale systems made up of a myriad of physical and virtual components. Microservice-based systems running on large Kubernetes clusters are a prime example. Components can fail in multiple ways. The premise of self-healing is that the overall system will not fail and will be able to automatically heal itself, even if this causes it to operate in a reduced capacity temporarily.

The building blocks of such reliable systems are as follows:

• Redundancy
• Observability
• Auto-recovery

The basic premise is that every component might fail – machines crash, disks get corrupted, network connections drop, configuration may get out of sync, new software releases have bugs, third-party ...