Differences

This shows you the differences between two versions of the page.

--- k8s:core:why_k8s [2026/05/31 02:29] – created phong2018
+++ k8s:core:why_k8s [2026/05/31 02:47] (current) – phong2018
@@ Line 1: / Line 1: @@
 ====== Why Kubernetes Exists ======
-Kubernetes exists to solve problems that appear when running containers in production.
+Kubernetes exists to solve **real production problems** when running containers at scale.
+---
 ===== 1. Manual container management is hard =====
-If you run containers manually:
-  * You must restart failed containers
-  * You must handle scaling yourself
-  * You must manage networking manually
-This becomes impossible at scale.
+Example:
+You run 10 Docker containers manually.
+Problems:
+  * One crashes → you must restart it
+  * Traffic increases → you must add more containers manually
+  * Servers fail → everything breaks
+👉 This does NOT scale.
+---
+===== 2. Need for self-healing =====
+Example:
+If your app runs 3 containers:
+App Pod 1 → OK
+App Pod 2 → CRASHED
+App Pod 3 → OK
-===== 2. Need for self-healing systems =====
-In production:
-  * Containers crash
-  * Servers fail
-  * Traffic increases suddenly
 Kubernetes automatically:
-  * Restarts failed containers
+  * Detects Pod 2 failure
-  * Replaces unhealthy pods
+  * Removes it
-  * Reschedules workloads to healthy nodes
+  * Creates a new Pod
+👉 No human needed
+---
 ===== 3. Need for scaling =====
-Traffic is not constant:
-  * Peak hours → high traffic
-  * Off-peak → low traffic
-Kubernetes supports:
+Example:
-  * Horizontal scaling (add/remove pods automatically)
-===== 4. Need for consistent deployment =====
+During peak traffic:
-Without Kubernetes:
-  * Deployments differ across environments
-With Kubernetes:
-  * Same YAML works everywhere (dev, staging, prod)
-===== 5. Cloud portability =====
+Normal: 3 pods
-Kubernetes works across:
+High traffic: 10 pods needed
-  * AWS
-  * GCP
-  * Azure
-  * On-prem servers
-So you avoid vendor lock-in.
+Kubernetes can:
+  * Add more pods automatically (autoscaling)
+  * Remove them when traffic decreases
+---
+===== 4. Need for reliability =====
+Example:
+If a server dies:
+Node 1 (dead)
+Node 2 (running)
+Node 3 (running)
+Kubernetes:
+  * Moves workloads from Node 1 to other nodes
+---
+===== 5. Need for consistency =====
+Example:
+Same app in:
+  * Dev
+  * Staging
+  * Production
+Kubernetes ensures:
+  * Same YAML works everywhere
+  * Same behavior across environments
+---
 ===== Summary =====
 Kubernetes exists to make systems:
-  * Automated
-  * Scalable
   * Reliable
-  * Portable
+  * Scalable
+  * Self-healing
+  * Automated