108 lines
3.1 KiB
Markdown
108 lines
3.1 KiB
Markdown
# Demo-Nginx Scaling History
|
||
|
||
This document tracks the scaling changes made to the demo-nginx deployment for capacity planning and audit purposes.
|
||
|
||
## Scaling Timeline
|
||
|
||
### 2026-01-12: Scale to 5 Replicas
|
||
**Change:** Increased from 3 to 5 replicas
|
||
**Reason:** Enhanced capacity and improved high availability
|
||
**Impact:**
|
||
- Additional 2 pod instances for better load distribution
|
||
- Improved resilience against node failures
|
||
- Better handling of traffic spikes
|
||
- Total resource allocation:
|
||
- CPU requests: 500m (5 × 100m)
|
||
- Memory requests: 640Mi (5 × 128Mi)
|
||
- CPU limits: 1000m (5 × 200m)
|
||
- Memory limits: 1280Mi (5 × 256Mi)
|
||
|
||
**Commit:** `0669ac66c8a651f74b4c16a4618db03bdd843c02`
|
||
**Deployment Method:** GitOps via ArgoCD automatic sync
|
||
|
||
### 2026-01-12: Scale to 3 Replicas
|
||
**Change:** Increased from 2 to 3 replicas
|
||
**Reason:** Initial scaling for improved availability
|
||
**Commit:** `d5ffa97159ec391de950dc2d861361074ff3eee8`
|
||
|
||
### Initial Deployment: 2 Replicas
|
||
**Change:** Initial deployment with 2 replicas
|
||
**Reason:** Baseline deployment for demo application
|
||
|
||
## Current Configuration
|
||
|
||
- **Replicas:** 5
|
||
- **Namespace:** demo-app
|
||
- **Image:** docker.io/vladcrypto/demo-nginx:main-50
|
||
- **Service Type:** ClusterIP
|
||
- **Port:** 80
|
||
- **Health Checks:** Liveness and Readiness probes enabled
|
||
- **Monitoring:** Prometheus scraping enabled
|
||
|
||
## Scaling Guidelines
|
||
|
||
### When to Scale Up
|
||
- CPU utilization consistently above 70%
|
||
- Memory utilization consistently above 75%
|
||
- Response time degradation observed
|
||
- Preparing for expected traffic increase
|
||
- Node maintenance requires pod redistribution
|
||
|
||
### When to Scale Down
|
||
- CPU utilization consistently below 30% for extended period
|
||
- Cost optimization requirements
|
||
- Application usage patterns indicate over-provisioning
|
||
|
||
## Resource Considerations
|
||
|
||
Each replica consumes:
|
||
- **CPU Request:** 100m
|
||
- **Memory Request:** 128Mi
|
||
- **CPU Limit:** 200m
|
||
- **Memory Limit:** 256Mi
|
||
|
||
Total cluster resources for 5 replicas:
|
||
- **Total CPU Requests:** 500m (0.5 cores)
|
||
- **Total Memory Requests:** 640Mi (~0.625 GB)
|
||
- **Total CPU Limits:** 1000m (1 core)
|
||
- **Total Memory Limits:** 1280Mi (~1.25 GB)
|
||
|
||
## Monitoring and Alerts
|
||
|
||
Monitor the following metrics in Grafana:
|
||
- Pod CPU/Memory utilization
|
||
- Request latency
|
||
- Error rates
|
||
- Pod restart count
|
||
- Service availability
|
||
|
||
## Rollback Procedure
|
||
|
||
If issues arise after scaling:
|
||
|
||
1. Revert the deployment.yaml in Git:
|
||
```bash
|
||
git revert <commit-hash>
|
||
git push
|
||
```
|
||
|
||
2. ArgoCD will automatically sync the change
|
||
|
||
3. Or manually scale via kubectl:
|
||
```bash
|
||
kubectl scale deployment demo-nginx -n demo-app --replicas=<previous-count>
|
||
```
|
||
|
||
## Related Documentation
|
||
|
||
- [CICD_GUIDE.md](./CICD_GUIDE.md) - CI/CD pipeline documentation
|
||
- [ROLLBACK_MANUAL.md](./ROLLBACK_MANUAL.md) - Detailed rollback procedures
|
||
- [README.md](./README.md) - General application documentation
|
||
|
||
## Notes
|
||
|
||
- All scaling operations are tracked in Git for audit trail
|
||
- ArgoCD manages automatic synchronization from Git to cluster
|
||
- Changes follow GitOps principles for declarative infrastructure
|
||
- Scaling decisions should be data-driven based on monitoring metrics
|