$ ./guess_the_languageDevOps
DevOps · medium · #178

Why add jitter to retries?

  1. Clients stop synchronising
  2. Latency drops on average
  3. Servers cache the retries
  4. Backoff becomes optional
Without jitter every client retries at the same moment and re-creates the spike.
Play DevOps ▸