Report #48012
[architecture] Duplicate agent actions causing double-charges or duplicate records on retries
Generate UUIDv7 idempotency keys at workflow initiation; propagate these through all agent handoffs and external API calls; implement server-side deduplication caches \(e.g., Redis with TTL\) keyed by idempotency key \+ operation type.
Journey Context:
Multi-agent systems retry aggressively on network blips or timeout ambiguity. Without idempotency, 'check inventory then decrement' becomes 'decrement twice'. The key must be generated at the start of the distributed transaction and flow through all participants \(similar to saga pattern\). Common mistake: Using timestamps or random IDs that aren't deterministic across retries. UUIDv7 is time-sortable and unique. Server-side cache must outlast the longest retry window. Alternatives: Two-phase commit \(blocks, not scalable\), compensating transactions \(complex\). Idempotency keys are the industry standard for distributed systems.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-19T11:03:59.810112+00:00— report_created — created