rrezartprebreza/spring-boot-skills / skills/spring-boot-3/idempotency-patterns
idempotency-patterns
VerifiedUse when making retried HTTP commands or message processing safe against duplicate effects, including database-backed request keys, payload conflicts and concurrent retries. Do not apply caching as a substitute for business idempotency.
Idempotency Patterns
Spring Boot 3 baseline
The examples use Java 17 and Jakarta APIs supported by this Boot version. Keep dependencies managed by the project's Boot BOM.
Define the contract
Identify the authenticated tenant/principal, operation, idempotency key, canonical request hash, retention window and replayable result. Scope uniqueness by principal/tenant and operation; a raw client key must never allow reading another user's result. Authenticate and authorize before both initial execution and replay. Bound key length and stored response size. Document what same-key/different-payload and still-in-progress requests return.
Database-local effects
For PostgreSQL, the schema and claim example uses a composite primary key and INSERT ON CONFLICT DO NOTHING RETURNING. In a single database transaction: claim the key, apply the business write, store the resulting status/body/selected headers, then commit. Roll back the claim with the business mutation on failure.
If the insert returns no row, load the existing result in a subsequent statement under READ COMMITTED, compare the canonical hash and replay only when it matches. A concurrent insert waits on the uniqueness constraint; bound lock waits and return a documented retryable outcome on timeout. With snapshot isolation, serialization failures require retrying the whole transaction. Do not catch a unique-constraint violation and continue in an already-aborted transaction.
A result row and business write must use the same transaction manager and database. Return the stored result rather than reconstructing it from mutable current entity state. Keep replay headers allowlisted; do not persist cookies, bearer tokens or hop-by-hop headers.
External effects and messages
A database transaction cannot atomically include an HTTP payment or email. Write an outbox record with the business mutation and let a retryable worker perform the external effect using the provider's idempotency support. Reconcile uncertain outcomes before retrying irreversible effects. For messages, record a unique consumer/event pair and apply the projection in one transaction; acknowledge only after commit. Retry failed transactions and route poison messages deliberately.
Define retention from the real client retry window. Deleting a key permits a later repeat to execute again; do not present TTL as an exactly-once guarantee. Protect stored responses as application data and delete them according to the project's retention policy.
Verification
Test simultaneous identical requests, conflicting payloads, tenant separation, rollback after claim, response replay after a lost connection and retries following a worker crash. Use the production database engine for lock and isolation tests, not H2 compatibility mode. The bad example illustrates the check-then-act race.
Gotchas
- Agent checks existence then inserts - enforce uniqueness in the database.
- Agent uses a JVM map or Redis TTL as the only business guard - coordinate with the actual write transaction.
- Agent replays another tenant's result - scope the key and re-authorize.
- Agent retries an external effect after a timeout blindly - reconcile or use provider idempotency.
- Agent deletes keys too early - document the retry window and expiry semantics.
Official sources
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| deprecation-and-migrationaddyosmani/agent-skillsManages deprecation and migration. Use when removing old systems, APIs, or features. Use when migrating users from one implementation to another. Use when migrating a database schema in production, such as renaming or dropping a column without downtime (expand/contract). Use when deciding whether to maintain or sunset existing code. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |
| performance-optimizationaddyosmani/agent-skillsOptimizes application performance across frontend, backend, queries, and databases. Use when performance requirements exist, when you suspect performance regressions, when Core Web Vitals or load times need improvement, when N+1 query patterns need fixing, or when profiling reveals bottlenecks. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |
| Agent Cost Reportthedotmack/claude-memBelievable agent cost report for any period, default the last 7 full days PT, not counting today. Measured tokens from Claude Code transcripts priced at OpenRouter list prices (ESTIMATED), measured provider spend when a sanctioned source exists, note-taker cost separate, Timing-style HTML/PDF plus report.json, line-items.csv, evidence.json. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |
| Agent Cost Reportthedotmack/claude-memBelievable agent cost report for any period, default the last 7 full days PT, not counting today. Measured tokens from Claude Code transcripts priced at OpenRouter list prices (ESTIMATED), measured provider spend when a sanctioned source exists, note-taker cost separate, Timing-style HTML/PDF plus report.json, line-items.csv, evidence.json. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |