Recent AWS and Google outages show that hyperscaler SLAs offer billing credits rather than operational remedies. Engineers must understand what downtime actually costs and plan for failure beyond the credit. By Michael Wallace.
Both put the work on the customer. AWS requires a claim through its Support Center by the end of the second billing cycle after the incident, with your own request logs documenting the errors. Google requires you to notify support within 60 days, with log files showing the downtime. Miss the deadline or the evidence, and you forfeit the credit.
Source: https://www.innoscale.net/cloud-outages-sla-credits/
Two recent cloud outages illustrate a critical gap between availability metrics and operational reality. AWS experienced a 5.5-hour incident in a Tokyo availability zone that blocked launches, scaling, and provisioning while leaving running instances unaffected. Google’s London outage lasted six hours after a sensing error prevented diesel generators from starting during a utility failure. Both events demonstrate that even hyperscalers fail, and the status signals engineers rely on can themselves be unreliable.
The remedy in both cases is a service credit, not a fix. AWS EC2 credits range from 10% to 100% of the monthly bill based on uptime shortfalls, while Google’s Compute Engine offers similar tiered credits. Both require customer claims with log evidence within strict deadlines, and both define the credit as the sole and exclusive remedy. Critically, AWS counts an instance as unavailable only when it lacks external connectivity. During the Tokyo incident, running instances stayed connected, so the inability to launch or scale may not have triggered the SLA at all.
For practitioners, the implication is clear: a credit does not cover lost revenue, missed deployments, or customer impact. Teams should calculate the actual cost of an hour of downtime and compare it against the SLA credit. The real question is who answers when things break, what telemetry they can access, and whether your architecture can survive a zone loss without relying on provider remediation. Good read!
[Read More]