Attributing causality to program execution capacity modifications

A method for a configured computing system of a program execution service to determine causality for dynamic modifications in program execution capacity provided to users, the method comprising:

• under control of the configured computing system of the program execution service, the program execution service providing a plurality of computing nodes that are configurable to execute programs of a plurality of remote users in exchange for fees charged to the users, and for each of multiple of the users,
• receiving information from the user that specifies an initial desired quantity of computing nodes for use in executing an indicated program of the user and that specifies multiple quantity modification triggers for use in later initiating automated modifications to the quantity of computing nodes being provided for use by the user, each of the quantity modification triggers including one or more criteria for use in determining if the quantity modification trigger is satisfied and including a specified computing node quantity change to be requested if the quantity modification trigger is satisfied;
• automatically determining a group for the user of multiple of the plurality of computing nodes for use in executing the indicated program of the user, the multiple computing nodes being of the initial desired computing node quantity, and providing the multiple computing nodes for use by the user at a first time;
• in response to an instruction from the user, initiating execution of a copy of the indicated program on each of the computing nodes of the group;
• and
• automatically managing the computing nodes of the group during a period of time from the first time to a later second time, by:
• during the period of time, automatically monitoring the computing nodes of the group to determine performance metrics from operation of those computing nodes, and automatically determining that at least one of the quantity modification triggers specified by the user is satisfied based on the determined performance metrics;
• during the period of time, receiving one or more quantity change instructions from the user that each indicates a dynamically specified request for a quantity change in computing nodes of the group;
• automatically determining an actual computing node quantity for the second time that reflects a quantity of the computing nodes of the group that are actually available at the second time to execute the program for the user, the determined actual computing node quantity at the second time being distinct from the initial desired quantity of computing nodes at the first time based at least in part on multiple changes to availability of computing nodes of the group during the period of time, at least some of the multiple availability changes having one or more associated fees for which the user is charged;
• for each of at least one of the multiple changes to availability, automatically attributing a single cause of the availability change that is either one of the at least one quantity modification triggers that are determined to be satisfied or one of the one or more quantity change instructions from the user;
• for each of at least one other of the multiple availability changes for which a single cause is not automatically attributed, the at least one other availability changes being distinct from the at least one availability changes, automatically attributing a combination of multiple possible causes to the availability change that are independent of each other and that include one or more of the at least one quantity modification triggers that are determined to be satisfied and include at least one of the one or more quantity change instructions from the user;
• and
• providing indications of the attributed single cause for each of the at least one availability changes and of the attributed combination of multiple causes for each of the at least one other availability changes, so as to enable the associated fees for the at least some availability changes to be imputed to the user.