Steadykey is making idempotency deterministic. It canonicalizes any JSON payload (order-insensitive, handles Maps/Sets/BigInts) and hashes it to a stable key [1]. Markers drop into pluggable stores—Redis, Postgres, MySQL, MongoDB, SQLite, or an in-memory adapter—so workflows finish cleanly.
Deterministic Idempotency Keys – Steadykey’s core idea is that a single, stable key represents a payload across runs. You can opt in to storing the canonical payload, refresh TTLs, or clear markers when a workflow finishes [1]. These options emphasize durable, repeatable decisions in distributed tasks.
Pluggable Stores Across Engines – The system drops markers into multiple databases, letting teams mix and match across Redis, Postgres, MySQL, MongoDB, SQLite, and in-memory storage as needed [1]. That pluggability is the hook for cross-store reliability patterns without rewriting logic per store.
Feedback & Docs – The discussion centers on hashing semantics, the adapter API, and gaps in docs or examples, with room for refinements as adoption grows [1].
Related Debate – A separate post argues PostgreSQL deserves better than libpq, highlighting tooling questions that echo across storage choices [2].
Closing thought: the push toward uniform idempotency semantics across heterogeneous stores is surfacing calls for better tooling and client libraries that make cross-store patterns feel native.
References
Show HN: Steadykey – Deterministic Idempotency Keys with Pluggable Stores
Discusses Steadykey for stable idempotency keys, JSON canonicalization, and pluggable adapters across Redis, Postgres, MySQL, MongoDB, SQLite, and memory.
View sourcePostgreSQL deserves better than libpq
Argues PostgreSQL client library libpq is insufficient; suggests improvements or alternatives for PostgreSQL usability
View source