Cloud Monitoring Strategy Development
Optimization Actions
| TECHMART | COST OPTIMIZATION ACTION PLAN | ||||
| Requirement | Proposed action | How to validate | Expected trade-off | Status |
| Idle / overprovisioned instances | Monitor average and peak CPU; reduce minimum baseline capacity where safe | EC2 utilization, Auto Scaling history | Too-small baseline increases cold starts | Proposed |
| Reserved / Spot opportunities | Use reserved commitments for measured steady usage; Spot only for fault-tolerant jobs | Compute Savings Plans / Reserved utilization, interruption tolerance | Commitment risk / Spot interruption | Proposed |
| GPU underutilization | Review image-processing workflow and lower-cost CPU/on-demand options | GPU metrics, per-job completion time and cost | Potential performance impact | Proposed |
| Storage lifecycle | Apply S3 tiering/expiration to appropriate logs and archives | Storage Inventory, retention and retrieval needs | Retrieval charges / compliance retention | Proposed |
| Billing configurations | Enable budgets and tagging for environment, service, owner | Budget alerts, tagged Cost Explorer views | Additional administration | Proposed |
| Service limits | Review quotas for EC2, Auto Scaling and load balancing | Service Quotas console and peak traffic test | Headroom vs. cost | Proposed |
Case Facts
| TECHMART | COST OPTIMIZATION EVIDENCE | |||
| Issue | Case evidence | Implication | Source |
| Infrastructure cost increase | 34% increase Aug–Oct 2024 | Review idle capacity and spending drivers | UMGC Unit 6 Case Scenario, p. 1 |
| Idle reserved instances | Identified as a cost driver | Right-size and review commitments | UMGC Unit 6 Case Scenario, p. 1 |
| Underused GPU instances | Used for image processing | Evaluate less expensive processing options | UMGC Unit 6 Case Scenario, p. 1 |
| Traffic peak | 280% increase in 4 hours | Use scalable capacity for seasonal demand | UMGC Unit 6 Case Scenario, p. 1 |
| Cost figures | Underlying AWS invoices not supplied | Any dollars below are illustrative, not actual spending | Case scenario limitation |
| Assignment requirement | Evidence to provide | Status | Source |
| Cost optimization spreadsheet and summary | This workbook includes cost model and proposed actions | Prepared | Unit 6 Assignment, p. 3 |
| Evaluate idle capacity, commitments, storage, billing | See Actions tab | Prepared | Unit 6 Assignment, p. 3 |
Scenario Model
| TECHMART | ILLUSTRATIVE MONTHLY SCENARIO | ||||
| Planning exercise only: hypothetical monthly USD—not an AWS bill, Cost Explorer export, or verified pricing. | ||||
| Category | Illustrative current ($) | Illustrative proposed ($) | Change ($) | Rationale |
| EC2 compute / GPU | $1,900 | $1,250 | ($650) | Reduce idle instances; assess GPU right-sizing |
| Load balancing / API / CDN | $350 | $440 | $90 | Additional routing/CDN costs to support availability |
| RDS / backups / cache | $780 | $930 | $150 | Resilience and cache spending may rise |
| Monitoring / logs / transfer | $420 | $400 | ($20) | Improve retention, sampling and log filters |
| Storage / network / other | $550 | $420 | ($130) | Storage lifecycle and resource cleanup |
| TOTAL | $4,000 | $3,440 | ($560) | |
| Metric | Calculated result | |||
| Illustrative monthly savings | $560 | |||
| Illustrative savings rate | 14.0% | |||
| Evidence status | Case provides percentage increase only; model values are hypothetical | |||
| Citation | UMGC Unit 6 Case Scenario p.1; Draft planning assumptions from Danish_Eqbal_CloudProject.docx |