“Slack operates a global, multi-regional edge network, but most of our core computational infrastructure resides in multiple Availability Zones within a single region, us-east-1.”
One flaky network link in one availability zone degraded Slack for 48 minutes on June 30, 2021, then failed again that evening. The 99.99% SLA allows less than an hour of downtime a year. The fix took 1.5 years and amounts to a button that tells Envoy to stop sending traffic to the bad zone. All of it still lives in us-east-1.