High Performance Cloud-Native Microservices
summary
Hazelcast provides distributed caching for Kubernetes with four concrete patterns: embedded (Java JVM only), sidecar (multi-language), HTTP reverse-proxy, and client‑server; Helm installs a three‑member cluster by default and Hazelcast Cloud offers managed clusters. Cache entries use RAM for lower latency, backup‑count defaults to 1 (configurable up to 6 or zero), and Jet offers stream processing. Best practices include setting requests and limits equal for JVM pods, enabling pod anti‑affinity, zone‑aware backups, and exposing Prometheus metrics.
AI-written summary · about
chapters (8)
- 0:00Introduction and agenda
- 2:55Why use distributed caching
- 5:05Caching patterns explained
- 11:12Kubernetes caching architectures
- 22:29Deploying Hazelcast in Kubernetes
- 26:44Best practices for caching
- 32:22Zone-aware deployment benefits
- 33:08Q&A and summary
related talks
45mWhere is my cache? Architectural patterns for caching microservices by example
53mLeveraging Caching For Cloud Native Applications
05mStartup Showcase: Hazelcast
55mHazelcast - Hitchhiker’s Guide. Grzegorz Piwowarek. Java Fest 2020 Online
44mHow to speed up your application using JCache
45mEasy Scaling with Open Source Data Structures
38mHazelcast That Could: Getting Lightning-Fast Responses at Hundreds of RPS
37m5 Levels of High Availability: from Multi-instance to Hybrid Cloud