IntegrationArchitect (Hardest)chaos-engineeringresiliencetesting
Design integration chaos engineering tests for Salesforce failure modes
Real World Scenario
Integrations never tested for Salesforce 503 or limit exhaustion; first real outage exceeds RTO by 6 hours.
Expected Answer
• Chaos experiments: inject 503 latency timeout in staging middleware
• Simulate daily API limit exhaustion verify graceful degradation
• Game day quarterly cross-team Salesforce outage drill
• Validate DLQ replay compensation saga rollback paths
• Measure RTO RPO actual not documented aspiration
• Automated fault injection CI staging pipeline monthly
• Blameless post-game-day architecture improvements
Follow-Up Questions & Answers
Click to expand — each follow-up includes a direct, interview-ready answer
Direct answer: Chaos experiments: inject 503 latency timeout in staging middleware Also consider: Simulate daily API limit exhaustion verify graceful degradation In practice: Game day quarterly cross-team Salesforce outage drill Document the decision in an ADR and align with enterprise standards.
Architect Perspective
Chaos engineering proves resilience claims—untested runbooks fail first real incident.