Techniques for generating a configuration for electrically isolating fault domains in a data center

One or more non-transitory computer-readable media comprising instructions that, when executed by one or more hardware processors, cause performance of operations comprising:

• assigning a first fault domain to a first electrical zone of a data center comprising a first power supply and a second power supply configured to provide a first redundant electrical power supply for the first electrical zone,
• wherein the first fault domain comprises a first virtual location for distributing a first compute instance;
• wherein based on assigning the first fault domain to the first electrical zone, the first compute instance executes on a first set of hardware components that share a first single point of failure with respect to the first electrical zone;
• detecting, via an event stream, an event associated with a power sensor indicative of a power outage of the first power supply corresponding to the first electrical zone of the data center, wherein the power outage of the first power supply renders the first electrical zone unprotected by electrical power supply redundancy;
• determining, based at least on a fault domain configuration for the data center indicating that the first fault domain is assigned to the first electrical zone, that the power outage of the first power supply renders the first fault domain unprotected by electrical power supply redundancy;
• and
• based at least on the power outage rendering the first fault domain unprotected by electrical power supply redundancy, assigning the first fault domain to a second electrical zone of the data center, wherein the second electrical zone provides a second redundant electrical power supply to the first fault domain,
• wherein based on assigning the first fault domain to the second electrical zone, the first compute instance executes on a second set of hardware components that receive the second redundant electrical power supply of the second electrical zone.

A system, comprising:

• at least one device including a hardware processor;
• the system being configured to perform operations comprising:
• assigning a first fault domain to a first electrical zone of a data center comprising a first power supply and a second power supply configured to provide a first redundant electrical power supply for the first electrical zone,
• wherein the first fault domain comprises a first virtual location for distributing a first compute instance;
• wherein based on assigning the first fault domain to the first electrical zone, the first compute instance executes on a first set of hardware components that share a first single point of failure with respect to the first electrical zone;
• detecting, via an event stream, an event associated with a power sensor indicative of a power outage of the first power supply corresponding to the first electrical zone of the data center, wherein the power outage of the first power supply renders the first electrical zone unprotected by electrical power supply redundancy;
• determining, based at least on a fault domain configuration for the data center indicating that the first fault domain is assigned to the first electrical zone, that the power outage of the first power supply renders the first fault domain unprotected by electrical power supply redundancy;
• and
• based at least on the power outage rendering the first fault domain unprotected by electrical power supply redundancy, assigning the first fault domain to a second electrical zone of the data center, wherein the second electrical zone provides a second redundant electrical power supply to the first fault domain,
• wherein based on assigning the first fault domain to the second electrical zone, the first compute instance executes on a second set of hardware components that receive the second redundant electrical power supply of the second electrical zone.