Capstone
Cloud & DevOps
Cloud-Native CI/CD Platform
GitOps pipeline with Kubernetes, Helm and progressive delivery
- Kubernetes
- Docker
- Helm
- Argo CD
- +2
₹5,299₹8,999
4.8/74 reviews
Cloud & DevOps · Capstone
Event-Driven Autoscaling with KEDA
Workers that scale on queue depth, down to zero when the queue is empty
4-5 weeks build
Overview
CPU is a poor signal for a worker that is waiting on a queue. Here KEDA watches RabbitMQ queue length and sizes a Python worker deployment to match, scaling to zero when nothing is waiting and back up when a burst arrives. Load experiments vary burst size and scaler settings, and the report compares KEDA with a CPU-based HPA on the same jobs. It all runs on a local kind or minikube cluster.
What makes it stand up
- 01Scaling driven by queue depth rather than CPU
- 02Scale-to-zero when idle, with the cold-start delay measured
- 03Polling interval, cooldown and target per replica tuned by experiment
- 04Side-by-side comparison against a CPU-based HPA on the same load
- 05Kafka scaler variant included for students who prefer Kafka
Module breakdown
- 01
Workload
Producer API and Python workers consuming from RabbitMQ.
- 02
KEDA
ScaledObject with the RabbitMQ trigger and authentication.
- 03
Load experiments
Burst and steady scenarios driven by a k6 producer.
- 04
Metrics
Prometheus graphs of queue length, replicas and job latency.
- 05
Comparison
The same runs under an HPA, with the gaps explained.
After this, you will be able to
- Explain why queue length is a better scaling signal for workers
- Defend your scaler settings with measured experiment results
- Demo workers scaling from zero as a burst of jobs arrives
Technology stack
- Kubernetes
- KEDA
- RabbitMQ
- Python
- Prometheus
- k6
- Level
- Capstone
- Domain
- Cloud & DevOps
- Build time
- 4-5 weeks
- Difficulty
You receive
- Complete, commented source code
- 60-90 page project report (IEEE format)
- Editable presentation deck
- Architecture, ER and UML diagrams
- Local setup and deployment guide
- Sample dataset or seed data
- Viva question bank with answers
- One 45-minute walkthrough call
- 30 days of doubt-clearing support