Power aware load placement

1. A computer-implemented method comprising:

receiving a request by a user of a cloud provider network to launch a virtual machine instance, the request indicating a user-specified preference of an instance type of the virtual machine instance;

determining, by respective lineup power sense circuitry, a power utilization for each lineup of a plurality of lineups that each provide power to a corresponding plurality of racks of physical servers;

selecting a first lineup of the plurality of lineups for the virtual machine instance based on the power utilizations for the plurality of lineups indicating the first lineup is not approaching and not exceeding a power utilization threshold of the plurality of lineups and a second lineup of the plurality of lineups is approaching or exceeding the power utilization threshold of the plurality of lineups;

selecting a virtual machine slot for the virtual machine instance from a plurality of candidate virtual machine slots of the physical servers of the first lineup of the plurality of lineups based on the user-specified preference; and

executing the virtual machine instance by the virtual machine slot of a physical server of the first lineup.

2. The computer-implemented method of claim 1 , wherein the power utilization threshold of the first lineup is the same as the power utilization threshold of the second lineup.

3. The computer-implemented method of claim 1 , further comprising:

receiving a lineup topology that indicates the physical servers of each lineup of the plurality of lineups;

placing a first replica of a dataset on one or more physical servers of a first lineup of the plurality of lineups; and

placing a second replica of the dataset on one or more physical servers of a second lineup of the plurality of lineups based on the lineup topology.

4. A computer-implemented method comprising:

receiving a request to launch a virtualized computing resource, the request indicating one or more virtualized computing resource preferences;

determining, by respective power sense circuitry, a power utilization for each lineup of a plurality of lineups that each provide power to a corresponding plurality of racks of physical servers;

selecting a physical server from a first lineup of the plurality of lineups at which to launch the virtualized computing resource based on the one or more virtualized computing resource preferences and the power utilizations for the plurality of lineups indicating the first lineup is not approaching a power utilization threshold of the plurality of lineups and a second lineup of the plurality of lineups is approaching the power utilization threshold of the plurality of lineups; and

executing the virtualized computing resource by the physical server.

5. The computer-implemented method of claim 4 , wherein the power utilization threshold of the first lineup is the same as the power utilization threshold of the second lineup.

6. The computer-implemented method of claim 4 , further comprising changing the power utilization threshold of the plurality of lineups to a different power utilization threshold.

7. The computer-implemented method of claim 4 , further comprising:

receiving a lineup topology that indicates the physical servers of each lineup of the plurality of lineups;

placing a first replica of a dataset on one or more physical servers of a first lineup of the plurality of lineups; and

placing a second replica of the dataset on one or more physical servers of a second lineup of the plurality of lineups based on the lineup topology.

8. The computer-implemented method of claim 4 , further comprising:

receiving a second request to launch a second virtualized computing resource, the second request indicating one or more second virtualized computing resource preferences;

selecting a physical server from a lineup of the plurality of lineups at which to launch the second virtualized computing resource based on the one or more second virtualized computing resource preferences and the power utilizations for the plurality of lineups; and

executing the second virtualized computing resource by the physical server, wherein the selecting the physical servers at which to launch the virtualized computing resource and the second virtualized computing resource comprises selecting physical servers to equalize power utilization of each lineup of the plurality of lineups based on the power utilizations.

9. The computer-implemented method of claim 8 , wherein the power utilization of each lineup is normalized by the number of racks in that lineup.

10. The computer-implemented method of claim 8 , wherein the selecting the physical servers at which to launch the virtualized computing resource and the second virtualized computing resource comprises overriding selection of physical servers on different lineups to equalize power utilization of each lineup of the plurality of lineups based on the power utilizations, and instead selecting physical servers on a single lineup of the plurality of lineups at which to launch the virtualized computing resource and the second virtualized computing resource.

11. The computer-implemented method of claim 4 , wherein each lineup is coupled to a separate power transformer of a plurality of power transformers.

12. The computer-implemented method of claim 4 , wherein the determining the power utilization for each lineup of the plurality of lineups comprises determining an average power utilized by each lineup for a time period.

13. The computer-implemented method of claim 4 , further comprising moving one or more virtualized computing resources to a different lineup to rebalance power utilization of each lineup of the plurality of lineups.

14. The computer-implemented method of claim 4 , wherein the plurality of lineups comprises three or more lineups, and the selecting the first lineup comprises selecting the first lineup having lowest power utilization of the power utilizations of the plurality of lineups.

15. A system comprising:

a plurality of physical servers; and

a control plane service of a provider network implemented by one or more electronic devices, the control plane service including instructions that upon execution cause the control plane service to:

receive a request to launch a virtualized computing resource, the request indicating one or more virtualized computing resource preferences;

determine, by respective power sense circuitry, a power utilization for each lineup of a plurality of lineups that each provide power to a corresponding plurality of racks of a proper subset of the plurality of physical servers;

select a physical server from a first lineup of the plurality of lineups at which to launch the virtualized computing resource based on the one or more virtualized computing resource preferences and the power utilizations for the plurality of lineups indicating the first lineup is not approaching a power utilization threshold of the plurality of lineups and a second lineup of the plurality of lineups is approaching the power utilization threshold of the plurality of lineups; and

execute the virtualized computing resource by the physical server.

16. The system of claim 15 , wherein the power utilization threshold of the first lineup is the same as the power utilization threshold of the second lineup.

17. The system of claim 15 , wherein the instructions, upon execution, further cause the control plane service to change the power utilization threshold of the plurality of lineups to a different power utilization threshold.

18. The system of claim 15 , wherein the instructions, upon execution, further cause the control plane service to:

receive a lineup topology that indicates the physical servers of each lineup of the plurality of lineups;

place a first replica of a dataset on one or more physical servers of a first lineup of the plurality of lineups; and

place a second replica of the dataset on one or more physical servers of a second lineup of the plurality of lineups based on the lineup topology.

19. The system of claim 15 , wherein each lineup is coupled to a separate power transformer of a plurality of power transformers.

20. The system of claim 15 , wherein the determination of the power utilization for each lineup of the plurality of lineups comprises determining an average power utilized by each lineup for a time period.