Software Performance Testing and Engineering Course

Explore the difference between functional and non-functional testing, then dive into core performance testing and performance engineering concepts, including load handling, concurrency, and CI/CD integration.

Define performance engineering as building software with non-functional attributes—speed, throughput, scalability, reliability, and availability—considered throughout the software development life cycle, while performance testing validates these requirements before deployment.

Master performance testing by understanding the full tech stack across operating system, network, disk, web, app, and database layers, and anticipate unknown unknowns that affect performance.

Learn the five-phase performance testing life cycle, with phase one test planning, documenting the application under test scope, tasks, and non-functional requirements, and conducting a signed-off stakeholder walkthrough.

Master phase two test preparation tasks—licensing, open source options, firewall setup, and data masking—then phase three load scripts, mocks, test data, workload models, and best practices for scripts.

Execute tests, analyze results, and tune performance to meet the peak load through warmup, baseline, endurance, scalability, stress, and resilience tests, with thorough documentation for closure.

Learn to download and set up Apache JMeter on Windows, launch the JMeter UI, and install custom plugins via the plugin manager to access advanced thread groups and graphs.

Nagraj demonstrates building a geometer script to test microservices secured by an API gateway, generating a security token and calling view and update account APIs.

Learn to build a JMeter rest api test script, configure a http sampler with token retrieval, headers, and query params, then extract the access token with a regex extractor.

Develop JMeter REST API scripts to test view account details and update account details, including token generation, correlation, and header management.

Parameterize api requests by reading data from an external csv file with csv data set config. Replace hard-coded values with dynamic account ids and verify responses show status equals success.

Learn to script REST API tests with JMeter by adding delays, parallelizing requests, applying throughput timers, and using Little's law to estimate required threads.

Compare monolithic and microservice architectures to understand scalability, independent deployment, and targeted testing, then learn how microservices improve resource utilization and rapid feature delivery.

Analyze performance testing of microservices by testing each service individually with mocks and service virtualization, then validate auto scaling and high availability.

Create an ECS cluster with the EC2 launch type, define a task definition, and deploy a service behind an application load balancer with security groups, port mappings, and health checks.

This lecture demonstrates scalability testing of a NodeJS microservice on AWS ECS using JMeter, evaluating view and update account details APIs at 30 tps with 95th percentile targets.

analyze scalability test results and changes to track improvements, plan a higher load toward 100 tps, and confirm 24 tps with sla-compliant response times.

Analyze the scalability results, showing 95th percentile response time under SLA and low error rate; conclude one ECS task handles 100 TPS and plan a two-task load at 60 TPS.

Configure autoscaling alarms at the ECS service level using CPU thresholds of 300% for scale up and 120% for scale down (40 percent of scale up), with a 30-second cooldown.

Configure auto scaling alarms at the ecs cluster level to automatically add or remove ec2 instances based on cpu and memory utilization, with min/max capacity and thresholds.

Explore a case study of batch file processing on AWS lambda, detailing data pull, validation, processing, and file generation for vendor delivery, and performance testing of 15,000 records per batch.

Explore a data processing case study where AWS Lambda consumes Kinesis events, tuning shards, concurrent batches per shard, and reserved concurrency to meet 20 TPS with minimal iterator age.

Explains why performance testing matters for cloud-hosted apps, outlining the shared responsibility model and auto-scaling metrics, with practical guidance for Easiest, Lambda, Kinases, and Dynamo DB.

Compare a performance tester with a performance engineer, and explore how load testing and cloud-based autoscaling shape performance. Learn about LoadRunner, JMeter, and monitoring tools.