Help center alerts by using metadata and offering multiple alert notification channels

摘要

A system and machine-implemented method relating to identifying anomalous events by estimating a processing time for an operation, estimating a processing time for an operation; calculating a maximum threshold time based on the estimated processing time for indicating anomalous processing of the operation, periodically determining during processing of the operation an amount of processing time used to perform the operation; and sending a notification to a user indicating an anomalous run of the operation has occurred if the determined amount of processing time exceeds the maximum threshold time.

主权项

1. A method for identifying anomalous events, the method comprising:
receiving data specifying a time window over which logged events are to be processed; determining, based on the specified time window and historical information, a quantity of logged events that are expected to be received during the time window; estimating, prior to a beginning of the time window and based on the quantity of logged events, a processing time for an operation that will process the logged events after the beginning of the time window; calculating a maximum threshold run time, based on the estimated processing time, for indicating an anomalous run time of the operation; determining, during processing of the operation, an amount of processing time used to perform the operation; and sending a first notification to a user, the first notification indicating an anomalous run time of the operation has occurred, if the determined amount of processing time exceeds the maximum threshold run time.