Chaos Report Template

Use this template for each multi-host chaos run package.

1. Run Metadata

• run_id:
• phase: small / medium / large
• chaos_profile: none / single_fault / combo_fault
• case_schema:
• global_store_profile: fast_failover / slow_cleanup
• orchestratorctl_meta:

2. Summary Gate Checklist

• all_get_complete:
• source_cardinality_timeline present:
• recover_time_sec:
• comm_bytes_delta:
• comm_errors_delta:
• retry_reason_buckets:
• budget_exit_reason_buckets:
• expected_failure_pass:
• failure_classification_counts:

3. Event Timeline

Attach events.jsonl and include a short readout:

• first chaos event timestamp:
• first request failure timestamp:
• recovery-complete timestamp:
• correlation ids used:

4. Budget Trace

Include aggregated budget fields and one example per case:

{
  "retry_reason_buckets": {
    "transport_timeout": 0
  },
  "budget_exit_reason_buckets": {
    "success": 0
  },
  "case_examples": [
    {
      "case_name": "example",
      "retry_reason_buckets": {},
      "budget_exit_reason_buckets": {}
    }
  ]
}

5. Failure Classification

Provide both aggregate counts and representative events:

{
  "infra": 0,
  "product": 0,
  "unknown": 0
}

6. Expected Failure Cases

For each negative case, record:

• expected_outcome=failure
• expected_error_pattern
• actual return code
• matched/not matched
• final status (expected_failure_pass / unexpected_failure / unexpected_success)

7. Handoff Attachments

• summary.json
• events.jsonl
• cases/<case_name>/result.json
• cases/<case_name>/metrics.json
• cases/<case_name>/classification.json
• meta/orchestratorctl_steps.jsonl