Self-tuning statistical resource leak detection

A computer-readable non-transitory storage medium configured with data and with instructions that when executed by at least one processor causes the processor(s) to perform a computational proactive process for self-tuned allocated resource usage leak detection, the process comprising the steps of:

• noting multiple resource usage sample points which show usage of an allocated resource, each noted sample point including a sample time and an allocated resource usage size, the noted sample points having a mean;
• tuning at least some of the noted sample points for allocated resource usage leak detection by computationally proactively utilizing at least one of the following self-tuning mechanisms:
• (a) exclusion of a noted sample point that is not within a specified standard deviation distance of the mean, (b) performance of a second order linear regression based indirectly on noted sample points, (c) identification of a derivative of a sequence of linear regression slopes;
• and
• statistically analyzing at least one result of the tuning step by proactively computing a linear regression slope and then using the slope in determining whether noted resource usage sizes have a trend.

A computer system equipped with self-tuning resource usage leak detection, the system comprising:

• a logical processor;
• a memory in operable communication with the logical processor;
• a plurality of resource usage sample points residing in the memory, each sample point including a sample time and an allocated resource usage size;
• a self-tuning resource usage leak detection code residing in the memory, which upon execution by the processor performs statistical trend analysis on at least a portion of the resource usage sample points;
• and
• wherein the system also comprises at least one of the following:
• (a) outlier resource usage sample points excluded from trend analysis by the leak detection code, (b) a plurality of linear regression slopes based on at least some of the resource usage sample points, as well as a second order linear regression that is based on the plurality of linear regression slopes, (c) a plurality of linear regression slopes based on at least some of the resource usage sample points, as well as a derivative of the plurality of linear regression slopes.