Implementation README

This repository is for the case-study of my master's thesis "A Lightweight Experimental Approach to Measure the Antifragility of Cloud-based Systems".

Structure

  • Setup_Notes.md: Notes on how to setup the experimental environment and helpful commands within kubectl, helm and litmus
  • Processing_Pipeline:
    • Jupyter Notebook Processing_Pipeline.ipynb used for calculating the antifragility score for the bank of anthos.
    • increase_load_*.txt: Log Output from Increase Load Experiment (used to extract timestamps)
    • perform_update_*.txt: Log Output from Perform Update Experiment (used to extract timestamps)
    • experiments_*.csv: Contain meta information for the litmus chaos experiments
    • inbound_traffic_*.csv: POST /login outbound latency within 2000ms
    • outbound_traffic_*.csv: POST /login inbound latency within 2000ms
    • login_requests_*.csv: POST /login requests within 2000ms latency and overall POST /login requests
    • slo_*.csv: "Service Level Objectives", percentage of requests within predefined latency
    • *.pdf and *.svg: Exported Graphs from the Jupyter Notebook
  • Chaos Scenarios:
    • Litmus Experiments
    • 10 Trials Experiment: Experiments that were actually executed and analyzed for the case study
  • Experiment Results:
    • Results of the Experiments
  • Prometheus: Dashboards for Grafana
  • linkerd: cluser certificates
  • bank-of-anthos: Cloned repo of Google's Bank of Anthos, but not modified
  • litmus: litmus helm chart, but not modified