Exploring High Availability in Kubernetes with Vitess

on YouTube ↗35m

" summary

Vitess provides high availability on Kubernetes by combining a stateless vtgate-as-proxy with stateful vt tablet guardians that do connection pooling, hot‑row protection, query/transaction timeouts and native backup/restore. Demo runs a truly global keyspace with four shards, masters and replicas across US‑East/West and Asia, uses Orchestrator for automated reparenting and shows killing a master and an entire data center: the cluster buffers, replans and self‑heals with brief errors only. Supports split/merge resharding online; MySQL first, Postgres support targeted in ~1–1.5 years.

AI-written summary · about

" chapters (4)

1. 0:00Start
2. 9:43Demo
3. 18:20Orchestrator
4. 29:05Add or Remove Shards

" related talks

• 35mExploring High Availability in Kubernetes with VitessJiten Vaidya · KubeCon + CloudNativeCon China 2019
• 45mDeep Dive: VitessDan Kozlowski, Jiten Vaidya · KubeCon + CloudNativeCon Europe 2019
• 31mVitess: Running MySQL with confidence on KubernetesSebastian Saemann · stackconf 2022
• 34mInside Vitess: Building Resilient, Distributed DatabasesHarshit Gangal · KubeCon + CloudNativeCon India 2025
• 43mVitess and a Kubernetes OperatorSugu Sougoumarane · KubeCon + CloudNativeCon Europe 2018
• 45mVitess: Stateless Storage in the CloudSugu Sougoumarane · KubeCon + CloudNativeCon North America 2019
• 44mIntro: VitessJiten Vaidya, Toliver Jue · KubeCon + CloudNativeCon China 2019
• 12mKeynote: Keeping Your Data in Kubernetes with VitessSugu Sougoumarane · Kubernetes Forum Bengaluru 2020