Performance Documentation

Abstract

This documentation section aims to introduce ideas and best practices on how OpenStack cloud may be exercised to evaluate its performance, what can be called performance at all and how it can be measured. Here you will find both testing/probing methodologies and test plans, proposed by OpenStack community members (as well as results of their execution against various cloud topologies).

Contents

• 1. Introduction
  • 1.1. What is the reason for writing this?
  • 1.2. Who will be interested in this?
  • 1.3. Who is contributing to this guide?
• 2. OpenStack summits news
  • 2.1. Ocata OpenStack summit recap
  • 2.2. Newton OpenStack summit recap
• 3. Methodologies
  • 3.1. Tools
  • 3.2. Methodology for testing Hyper-Scale
  • 3.3. Methodology for Containerized Openstack Monitoring
  • 3.4. HA InfluxDB as an external storage for Prometheus
• 4. Labs
  • 4.1. Intel-Mirantis Performance-Team Lab #1
  • 4.2. Intel-Mirantis Performance-Team Lab #2
  • 4.3. Grid’5000
• 5. Test Plans
  • 5.1. 1000 Compute nodes resource consumption/scalability testing
  • 5.2. Measuring performance of container repositories
  • 5.3. OpenStack control plane performance test plan
  • 5.4. OpenStack control plane density testing
  • 5.5. SQL Database Test Plan
  • 5.6. Fuel Containerized Control Plane performance testing
  • 5.7. Fuel Containerized Control Plane upgradability performance testing
  • 5.8. Measuring performance of Cinder with Ceph backend
  • 5.9. Keystone Performance testing
  • 5.10. Kubernetes density testing
  • 5.11. Kubernetes DNS performance test plan
  • 5.12. Kubernetes POD recovery test plan
  • 5.13. Kubernetes proxy performance test plan
  • 5.14. Massively Distributed RPCs
  • 5.15. Message Queue Performance
  • 5.16. Message Queue High Availability
  • 5.17. OpenStack Networking (Neutron) control plane performance test plan
  • 5.18. OpenStack Neutron Testing During Open Stack Upgrade Test Plan
  • 5.19. Neutron ML2/ODL Performance testing
  • 5.20. OpenStack API Performance Metrics
  • 5.21. OpenStack API worker performance and scalability testing with Browbeat
  • 5.22. OpenStack load testing
  • 5.23. OpenStack in a WAN Context Testing
  • 5.24. Measuring performance of provisioning systems
  • 5.25. Telemetry Services resource consumption/scalability testing
  • 5.26. OpenStack tenant networking performance
  • 5.27. Ceph RBD performance testing
  • Container cluster systems
  • 5.28. Hardware features test plans
  • 5.29. Neutron features test plans
  • 5.30. OpenStack reliability testing
• 6. Test Results
  • 6.1. 1000 Compute nodes resource scalability testing
  • 6.2. Ceph RBD performance report
  • Container cluster systems performance
  • Container repositories performance
  • 6.3. Results of OpenStack Services density testing
  • 6.4. Openstack control plane performance tests results
  • 6.5. SQL Database performance report
  • 6.6. SQL Database on Kubernetes performance report
  • 6.7. Fuel Containerized Control Plane performance report
  • 6.8. Fuel Containerized Control Plane performance report during upgrade
  • 6.9. Hardware features testing
  • 6.10. Results of measuring performance of Cinder with Ceph backend
  • 6.11. Keystone Performance Test Reports
  • 6.12. Massively Distributed RPCs report
  • 6.13. Environment Description
  • 6.14. Monitoring systems test results
  • 6.15. Message Queue Performance Reports
  • 6.16. RabbitMQ HA test reports
  • 6.17. OpenStack Networking (Neutron) control plane performance report for 400 nodes
  • 6.18. Neutron features scale testing
  • 6.19. Fuel Containerized Control Plane performance report during upgrade
  • 6.20. OpenStack load testing report
  • Provisioning systems test results
  • 6.21. OpenStack reliability testing
  • 6.22. Telemetry Services resource consumption/scalability test results on Gnocchi with CephStorage driver
  • 6.23. OpenStack tenant networking
• 7. Issues Analysis
  • 7.1. Kubernetes Issues At Scale 900 Minions