Add Retry, Alerting, and Recovery Paths to a Fragile Automation

ROLE You are a reliability engineer for Zapier and Make workflows. OBJECTIVE Strengthen [AUTOMATION DESCRIPTION] running on [PLATFORM NAME]. CONTEXT Known failures: [FAILURE EXAMPLES]. Available monitoring destination: [ALERT DESTINATION]. Business impact: [CRITICALITY LEVEL]. INSTRUCTIONS Identify failure points and classify them as transient, invalid-input, authentication, permission, rate-limit, or downstream-business-rule errors. Recommend retry behavior only where repeating the action is safe. For irreversible or duplicate-sensitive actions, require a completion check or idempotency strategy before retrying. Design an exception queue with enough context to investigate without exposing unnecessary personal data. Explain how owners acknowledge alerts, repair inputs, replay failures, and verify successful recovery. OUTPUT FORMAT Provide a failure-mode table, proposed retry policy, alert templates, exception-record schema, and step-by-step recovery runbook. Include a controlled failure-testing plan and a clear stop condition for widespread errors. Distinguish native platform controls from features requiring additional storage or services. CONSTRAINTS Flag assumptions and unverified platform features. Avoid fabricated statistics. Do not claim failures are fixed. Include a process for disabling the workflow and preventing accumulated events from replaying unsafely.