Autonomous application-level auto-scaling in a cloud

A system comprising:

• a memory;
• at least one processor coupled to the memory;
• a data collector to collect, using the at least one processor, at a virtual instance of a machine provided within a virtualization service, application-level data, the application-level data reflecting demand for a subject computing application executing on the virtual instance of the machine, wherein the subject computing application executes on a plurality of virtual instances each having a corresponding scaling module that performs scaling decisions;
• and
• a scaling module within the virtual instance of the machine to:
• select a scaling action based on the application-level data,
• lock a database row within a database, the database row associated with the subject computing application;
• verify that another scaling module executing at another virtual instance within the plurality of virtual instances has not performed the scaling action within a predetermined time period based on information stored in the database, and
• issue a request to perform the scaling action with respect to the virtual instance of the machine, the scaling action comprising terminating the virtual instance of the machine or instantiating a new virtual instance of the machine, the new virtual instance of the machine executing the subject computing application.

A machine-readable non-transitory storage medium having instruction data to cause a machine to:

• collect, using the at least one processor, at a virtual instance of a machine provided within a virtualization service, application-level data, the application-level data reflecting demand for a subject computing application executing on the virtual instance of the machine, wherein the subject computing application executes on a plurality of virtual instances each having a corresponding scaling module that performs scaling decisions;
• select, at the virtual instance, a scaling action based on the application-level data;
• locking a database row within a database, the database row associated with the subject computing application;
• verify that another scaling module executing at another virtual instance within the plurality of virtual instances has not performed the scaling action within a predetermined time period based on information stored in the database;
• and
• issue a request to perform the scaling action with respect to the virtual instance of the machine, the scaling action comprising terminating the virtual instance of the machine or instantiating a new virtual instance of the machine, the new virtual instance of the machine executing the subject computing application.