3 ms·

> The outage was not caused by a failure to code for the leap year. We had instability in a part of our infrastructure that allows our systems to communicate wi

> The outage was not caused by a failure to code for the leap year. We had instability in a part of our infrastructure that allows our systems to communicate with each other. https://twitter.com/askrobinhood/status/1234861941413351434?s=21 https://twitter.com/askrobinhood/status/1234861941413351434?... Sounds like Kafka / message bus to me

yushuf 7y ago
There Kafka cluster is actually managed by Confluent. I wonder if that has anything to do with the persistent outage
conduit242 7y ago
Could easily be service discovery too, like Zookeeper