Techniques for managing performance metrics associated with software applications

A method for implementing differential privacy in application performance metric reporting, the method comprising, by a computing device:

• receiving a plurality of application performance metric values associated with a peer group of software applications;
• generating a noise-adjusted percentile value by performing a differential privacy operation that includes applying noise to eliminate duplicate values within the plurality of application performance metric values;
• and
• presenting a visual representation of the noise-adjusted percentile value in a user interface.

A system comprising:

• a processor;
• a memory storing instructions that, when executed processor, cause the system to:
• receive a plurality of application performance metric values associated with a peer group of software applications;
• generate a noise-adjusted percentile value by performing a differential privacy operation that includes applying noise to eliminate duplicate values within the plurality of application performance metric values;
• and
• present, on a display coupled to the processor, a visual representation of the noise-adjusted percentile value in a user interface.

A non-transitory computer-readable medium storing instructions that, when executed by a processor of a server computing device, cause the server computing device to carry out operations for implementing differential privacy in application performance metric reporting, wherein the operations comprise:

• receiving, at the server computing device, a plurality of application performance metric values associated with a peer group of software applications;
• adding random noise to each of the plurality of application performance metric values to generate noise-adjusted application performance metric values, wherein the random noise is of a magnitude smaller than a display precision used for reporting percentile values, such that duplicate values within the plurality of application performance metric values are eliminated;
• sorting the noise-adjusted application performance metric values;
• generating a noise-adjusted percentile value for a target percentile by performing a differential privacy operation that selects the noise-adjusted percentile value from an interval between adjacent noise-adjusted application performance metric values;
• and
• causing a user interface to present a visual representation of the noise-adjusted percentile value without revealing underlying individual application performance metric values of other software applications in the peer group.