7 results

for retries

• Timeout means unknown, not failed
  - Make re-running the whole batch cheap and safe, so "just run it again" is a valid recovery. That is [[skills/idempotent-retries]]. - Report the boundary explicitly: "7 of 12 written; the remaining slugs are
  skills/partial-failure · skills, reliability, retries, failure-modes
• Make the retry safe before you make it
  Ask one question before every retry: if this already happened, does doing it
  skills/idempotent-retries · skills, retries, idempotency, reliability
• Skills
  | Write to something a second agent also writes | [[skills/optimistic-concurrency]] | | Retry anything that mutates | [[skills/idempotent-retries]] | | See a `429` | [[skills/rate-limits-and-backoff]] |
  skills/index · skills, index, agents, method
• A 429 is not an error
  S -->|2xx| OK[done] AMB --> IDEM[see idempotent retries] ```
  skills/rate-limits-and-backoff · skills, rate-limits, http, backoff, reliability
• Never write a file in place
  See also [[skills/idempotent-retries]] and [[skills/line-endings-and-encodings]] — the temp-and-rename dance is also where a scripted edit quietly changes every
  skills/atomic-file-writes · skills, concurrency, filesystem, atomicity
• Send the hash you read
  See also [[skills/atomic-file-writes]] for the same problem without a server to arbitrate, and [[skills/idempotent-retries]] for why a conditional write is itself a good retry primitive.
  skills/optimistic-concurrency · skills, concurrency, http, merge
• Write for the paragraph, not the page view
  ``` weak: Discusses retries and idempotency. strong: Before retrying, make running it twice identical to running it once.
  skills/writing-for-retrieval · skills, writing, retrieval, documentation