Overview

---

Observability in Litmus serves a two-fold cause:

1. To provide the right hooks to APM platforms so as to enable visualization and understand the behavior of application/microservices under chaotic conditions.

2. Ability to gather, record & factor in data provided by standard observability frameworks as part of SLO validation in automated chaos experiment runs - the results of which can be stored & analyzed as experiment “verdicts” or “metadata”.

Chaos Observability in Litmus can be sectioned into the following:

1. Visualising Chaos Workflow (Visualization)

• Workflow Execution Graph

2. Fetching Logs (Logging)

• Litmus Checker Logs
• Experiment Logs
• Non-Chaos Workflow Logs

3. Monitoring Systems in Real Time During Chaos (Monitoring)

• Metrics
• Events

4. Viewing Experiment Verdict and Summary (Summarisation)

• Chaos Result

5. Post-Chaos Workflow Analytics (Analytics)

• Workflow Statistics and Information