April 2018
Intermediate to advanced
468 pages
14h 34m
English
In this section, we will look at how Kubernetes can help operate a huge platform such as Hue. Kubernetes itself provides a lot of capabilities to orchestrate pods and manage quotas and limits, detecting and recovering from certain types of generic failures (hardware malfunctions, process crashes, and unreachable services). But, in a complicated system such as Hue, pods and services may be up and running but in an invalid state or waiting for other dependencies in order to perform their duties. This is tricky because if a service or pod is not ready yet, but is already receiving requests, then you need to manage it somehow: fail (puts responsibility on the caller), retry (how many times? for how long? ...
Read now
Unlock full access