Retry transient Postgres failures

At a glanceCopy Anchor Link

• Who: Workers and job runners using @zedtech/zedkit-core backoff helpers.
• What: Map serialization/deadlock/connection errors to retry without duplicating SQLSTATE lists.
• Use when: After a failed query in a retry loop.
• Avoid when: For HTTP mapping — use pgErrorCode and your API error model.
• See also: Tell a Postgres conflict from a bug

InstallCopy Anchor Link

pnpm add @zedtech/zedkit-server

Latest: 0.2.1 (via changesets).

ExampleCopy Anchor Link

import { pgErrorRetryable } from "@zedtech/zedkit-server";
import { computeBackoffMs } from "@zedtech/zedkit-server";

if (pgErrorRetryable(error) === true) {
  await sleep(computeBackoffMs(attempt));
  continue;
}

Wraps: (none)