Shopify outage.
Clearly Shopify has at least one single point of failure in their architecture.
Also sounds like they are not feature flagging their code changes well (to enable progressive per-tenant enablement) and they aren’t implementing redundancy of services with rolling code change approach and non-breaking changes.
Yes, it is very hard to get this consistently right, but at least you can mitigate and avoid having to be so draconian with code freezes.
I suspect they have most of this, they are a pretty tight engineering organization. no one is perfect however, especially at the speed they innovate.