Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Neotys PAC - Bruno Da Silva

330 views

Published on

Neotys organized its first Performance Advisory Council in Scotland, the 14th & 15th of November.
With 15 Load Testing experts from several countries (UK, France, New-Zeland, Germany, USA, Australia, India…) we will explore several theme around Load Testing such as DevOps, Shift Right, AI etc.
By discussing around their experience, the methods they used, their data analysis and their interpretation, we will create a lot of high-value added content that you will be able to use to discover what will be the future of Load Testing.

Published in: Engineering
  • Be the first to comment

  • Be the first to like this

Neotys PAC - Bruno Da Silva

  1. 1. Strengthening your performance testing policy Bruno Da Silva
  2. 2. Performance101 Performance drivers : • application functionalities, design patterns, frameworks, backend usage… • resources clustering, virtualization, network capacity, shared components… • load use cases, user profiles, load peaks, new app vs existing app… How to improve Performance Testing? • shift left : Continuous testing • shift right : Improving with production feedback resources
  3. 3. Stage 1 : Load test design – Best practices
  4. 4. Stage1 : Load testdesign – Best practices • Continuous performance : testing application during development ensure performance from the start 100% use cases available for final load tests • Production feedback : enhance load test with real usage adjust use cases for more accurate tests and better load distribution validate performance stability for future releases • Monitor your environment baseline for your tests and production behaviors measure the impact on shared components
  5. 5. Stage1 : Load testdesign – Best practices Prediction is hard when estimating load, especially for new applications Get ready but know your limits ! • 2 axes for Performance testing : load testing : evaluate compliance with specified performance requirements stress testing : evaluate the system or one of its components beyond the limits specified
  6. 6. Stage 2 : Data generation strategy
  7. 7. Stage2 : Data generationstrategy • Data is essential to execute load tests. Typical contexts : First use : low loaded database and many creations Campaign use : high loaded database and many consultations and editions • Mixing user profiles : New user : measure the impact of a fresh visit Returning user : measure the impact on cache • Data generation is touchy : Validate your rules for data generation (persistent data and datasets) Create a reference (database dumps) for better analysis among tests
  8. 8. Stage2 : Data generationstrategy • Ready to load : use data “as is” for production-like testing create more datasets with reliable data • When generating files, make sure to vary number and size to identify performance issues VS • Generate dumps from Production data problem : forbidden by data protection laws • the key to data anonymization : identify and substitute personal data with fake automate substitution logic to save time
  9. 9. Final stage : DevOps for Performance
  10. 10. Final Stage: DevOps for Performance Category Goals Example Performance testing • Load and stress testing • Improve future tests with framework parameters Neoload Monitoring • Analyze your load tests performance issues • Compare your production cases with your tests Dynatrace Continuous performance • Capitalize on your load tests to avoid performance regression • Test and monitor automatically after each release Jenkins Deployments • Automate the application deployment for consistency • Scale easily by creating injectors instances on the fly Openshift Generating data • Capitalize on data you need for your load tests • Automate data generation and the environment test prior to your tests Mockaroo
  11. 11. Final Stage: DevOps for Performance Improve the application AND performance tests from experience Address performance from the beginning to reduce costs
  12. 12. Thank you ! Q & A

×