Container cluster systems performance

• 1. Results of measuring performance of Mesos Marathon
  • 1.1. Environment description
    • 1.1.1. Cluster description
    • 1.1.2. Software versions
    • 1.1.3. Running options
    • 1.1.4. Hardware configuration of each server
    • 1.1.5. Network scheme and part of configuration of hardware network switches
  • 1.2. Results
    • 1.2.1. Test create
    • 1.2.2. Test delete
    • 1.2.3. Test restart
    • 1.2.4. Test update_cpu
    • 1.2.5. Test update_disk
    • 1.2.6. Test update_instances
    • 1.2.7. Test update_mem
    • 1.2.8. Test health check
• 2. Kubernetes system performance
  • 2.1. Results of measuring of API performance of Kubernetes
    • 2.1.1. Environment description
      • 2.1.1.1. Hardware configuration of each server
      • 2.1.1.2. Network scheme and part of configuration of hardware network switches
      • 2.1.1.3. Software configuration of Kubernetes service
      • 2.1.1.4. Software configuration of Test tool:
    • 2.1.2. Testing process
      • 2.1.2.1. Preparation
    • 2.1.3. Results
      • 2.1.3.1. 10-nodes cluster (all values are presented in milliseconds)
      • 2.1.3.2. 50-nodes cluster (all values are presented in milliseconds)
      • 2.1.3.3. 355-nodes cluster (all values are presented in milliseconds)
      • 2.1.3.4. Comparation
      • 2.1.3.5. Kubernetes pod startup latency measurement
    • 2.1.4. Applications
      • 2.1.4.1. Files and scripts to build Docker container with e2e-test tool
      • 2.1.4.2. Files and scripts to run Docker container with e2e-test tool
      • 2.1.4.3. Script to convert JSON from log file to RST table
      • 2.1.4.4. Pod startup time measurement toolkit
  • 2.2. Kubernetes density test report
    • 2.2.1. Environment description
      • 2.2.1.1. Software
    • 2.2.2. Reports
      • 2.2.2.1. Test Case #1: Maximum pods per node
      • 2.2.2.2. Test Case #2: Measure Kubelet capacity
• 3. Results of measuring performance of Kargo
  • 3.1. Environment description
    • 3.1.1. Hardware configuration of each server
    • 3.1.2. Network scheme and part of configuration of hardware network switches
    • 3.1.3. Software configuration of Kargo
      • 3.1.3.1. Setting up Kargo:
      • 3.1.3.2. Test tool:
      • 3.1.3.3. Operating system configuration:
  • 3.2. Testing process
  • 3.3. Results
  • 3.4. Issues that have been found during the tests
  • 3.5. Applications
    • 3.5.1. Launcher script
    • 3.5.2. Script for parsing collected metrics
• 4. Results of the ETCD health tests
  • 4.1. Environment description
    • 4.1.1. Hardware configuration of each server
    • 4.1.2. Network scheme and part of configuration of hardware network switches
    • 4.1.3. Software configuration
      • 4.1.3.1. Setting up Kubernetes cluster:
      • 4.1.3.2. Test tools
  • 4.2. Testing process
  • 4.3. Results
    • 4.3.1. Case 1: “ETCD vs PODS”
      • 4.3.1.1. ETCD performance(time series graphs)
      • 4.3.1.2. ETCD performance(number of pods dependency)
    • 4.3.2. Case 2: “ETCD vs SERVICES”
      • 4.3.2.1. ETCD performance(time series graphs)
      • 4.3.2.2. ETCD performance(number of services dependency)
    • 4.3.3. Analysis
      • 4.3.3.1. Case 1 (ETCD vs PODS)
      • 4.3.3.2. Case 2 (ETCD vs SERVICES)
  • 4.4. Applications
    • 4.4.1. prepare-minion.py