system-design – Page 17 – Türker Yıldırım

Jitter matters more than the backoff curve

NoteSeptember 9, 20211 min read

Exponential backoff spreads a single client’s retries over time and does nothing about a thousand clients retrying in lockstep after one outage. The recovering-service failure is the one…

The certificate that expired on a root nobody controlled

Deep diveSeptember 7, 20219 min read

30 September, and a certificate chain that was valid on every browser we tested. The API works and three integrations do not.

A CSV export that streams instead of collapsing

Deep diveJuly 29, 202110 min read

An export that worked until the account with 400,000 rows. fetchAll, a collection and a string built in memory, at 512 MB.

An idempotency key comes from the client or not at all

NoteJune 17, 20211 min read

The server cannot tell two identical requests apart from one request sent twice, because the payload is the same either way — only the client knows which it…

A healthcheck that tests the app, not the port

NoteJune 15, 20212 min read

A health check that opens a TCP connection reports healthy for a process that is listening and cannot serve a request. The rule that has held up is…

Burn rate beats a threshold on the raw metric

NoteJune 10, 20212 min read

A threshold on an error rate fires at the same point whether the budget is fresh or already spent, which is not the decision anybody wants to make.…

Queue priority is not a number, it is a separate queue

Deep diveJune 8, 202110 min read

Password resets stuck behind a nightly export. A priority field does not help when the worker is already four minutes into something else.

The outbox pattern, because two writes are not atomic

NoteJune 3, 20212 min read

Writing a row and publishing a message are two systems, and any ordering of them has a window where one succeeded and the other did not. Both writes…

A webhook endpoint that survives the sender

Deep diveJune 2, 202110 min read

The payment provider retried for six hours and every attempt timed out. A webhook is a notification, not a request for processing.

Terraform 1.0 is a promise about the next five years

Deep diveJune 2, 202110 min read

A version number that is a compatibility statement rather than a feature list. Three environments described by hand, and the import that is most of the work.