Cost-efficient high-availability multi-single-tenant services

A system for managing pooled computing resources used for providing services in a distributed system, comprising:

• a primary pool of computing resources designated for running an individual service instance in an isolated environment;
• a secondary pool of computing resources designated for running one or more secondary jobs different than the individual service instance;
• and
• one or more processors configured to:
• detect a failure condition associated with the isolated environment;
• select at least one of the one or more secondary jobs to preempt, the one or more secondary jobs comprising preemptible jobs each having a lower priority than the individual service instance;
• free computing resources in the secondary pool by reducing the computing resources allocated to the selected at least one secondary job;
• and
• dynamically allocate the freed resources to the individual service instances of the isolated environment with the failure condition;
• wherein the computing resources in the secondary pool are freed in proportion to the dynamically allocated freed computing resources.

A non-transitory computer-readable medium storing instructions executable by one or more processors for performing a method of managing pooled computing resources used for providing services in a distributed system, the method comprising:

• determining a primary pool of computing resources designated for running an individual service instance within an isolated environment;
• determining a secondary pool of computing resources designated for running one or more secondary jobs different than the individual service instance, the one or more secondary jobs comprising preemptible jobs each having a lower priority than the individual service instance;
• detecting a failure condition associated with the isolated environment;
• selecting at least one of the one or more secondary jobs to preempt;
• and
• freeing the computing resources in the secondary pool by reducing the computing resources allocated to the selected at least one secondary job, while dynamically allocating the freed computing resources to the individual service instance of the isolated environment with the failure condition, wherein the freeing is performed in proportion to the dynamically allocating of the freed computing resources.