Generating pseudo heat maps

A computer-implemented method, comprising:

• receiving, via a processor, temperature data from a plurality of sensors disposed at temperature-sensitive locations in a plurality of server racks;
• generating, via the processor, a continuous spatial statistical distribution model based on the temperature data;
• and
• generating, via the processor, a pseudo heat map comprising a plurality of discrete locations represented by geometric shapes and emphasizing an attribute of the continuous spatial statistical distribution model, wherein the emphasized attribute comprises an order in which a subset of the plurality of server racks overheats;
• calculating, via the processor, an anomaly score for each of the plurality of discrete locations based on a temperature probability, wherein the anomaly score is calculated as an inverse logarithmic probability of a given temperature for a given rack at a given time;
• and
• detecting, via the processor, a thermal anomaly based on threshold anomaly score being exceeded at a discrete location, wherein the processor is to determine an order in which to shut down servers in the subset of the plurality of server racks at the discrete location or an order in which to enforce backups of the servers in the subset of the plurality of server racks at the discrete location based on the order in which the subset of the plurality of server racks overheat in the pseudo heat map.

A computer program product for detection of anomalies using a pseudo heat map, the computer program product comprising a computer-readable storage medium having program code embodied therewith, wherein the computer readable storage medium is not a transitory signal per se, the program code executable by a processor to cause the processor to:

• receive temperature data from a plurality of sensors disposed at temperature-sensitive points of a plurality of server racks;
• generate a continuous spatial statistical distribution model based on the temperature data;
• generate the pseudo heat map comprising a plurality of discrete locations represented by geometric shapes and based on the continuous spatial statistical distribution model, wherein the pseudo heat map is to emphasize an attribute of the continuous spatial statistical distribution model comprising an order in which a subset of the plurality of server racks overheat;
• calculate an anomaly score for each of the plurality of discrete locations based on a temperature probability, wherein the anomaly score is calculated as an inverse logarithmic probability of a given temperature for a given rack at a given time;
• detect a thermal anomaly based on a threshold anomaly score being exceeded at a server rack, wherein the processor is to determine an order in which to shut down servers in the subset of the plurality of server racks at a discrete location or an order in which to enforce backups of the servers in the subset of the plurality of server racks at the discrete location based on the order in which the subset of the plurality of server racks overheat in the pseudo heat map;
• and
• generate an alert based on the detected thermal anomaly and send a request to shut down a server located in the subset of the plurality of server racks.