Optimizing Kubernetes Operators and Admission Control with Pepr
summary
Pepr is a TypeScript framework that builds controllers and admission logic using TypeScript framework, Server Side Apply, and mutating and validating webhooks. It uses the Kubernetes Watch API, a reconcile queue that preserves event order, and a built‑in etcd‑backed store and scheduler to apply sane defaults, auto‑create networkpolicies/serviceMonitors, and redeploy deleted resources. Examples include rejecting pods running as root, blocking hostPath writes, auto‑mutating securityContext and resource limits, and producing lightweight operators for multi‑cluster and air‑gapped deployment workflows.
AI-written summary · about
chapters (15)
- 0:00Pepr and Kubernetes operators
- 4:16Sparky’s chaotic first week
- 8:00Admission controllers with bouncers
- 10:04Pepr’s API for admission control
- 13:14Enforcing security defaults
- 16:22Mutating vs validating webhooks
- 17:46Custom resources with Pepr
- 22:43Pepr’s reconcile loop
- 23:26Pepr’s vision and CNCF
- 25:13Q&A: Admission control phases
- 26:58Q&A: Watch vs admission control
- 29:28Q&A: Pepr use cases
- 32:38Q&A: Interesting Pepr use case
- 34:26Q&A: Students or teams with Pepr?
- 35:49Q&A: Pepr’s lightweight architecture
related talks
32mMinimalist Best Practices in Kubernetes Operators
11mK8s Resources Ensuring Performance and Optimization- 32mMinimalist Best Practices in Kubernetes Operators
01h 01mKubernetes – going beyond the basics
59mKubernetes - going beyond the basics
33mBringing Pod Security Admission to Your Kubernetes Distribution
52mKubernetes - going beyond the basics
25mDesign Patterns for Extendable, Scalable K8s Extensions