The techniques described herein enable the use of a time factor for traffic management and routing. A system is configured to analyze traffic for a service over a period of time and identify (e.g., learn) traffic patterns that reflect a substantial effect on traffic during a particular real-world event. Using the traffic patterns identified via the analysis, the system can provide valuable time-based traffic information to service providers. A service provider can then create a predefined time-based profile that is used by a traffic manager to switch from a current traffic routing configuration to a different traffic routing configuration that better accommodates an expected traffic load for various endpoints. The predefined time-based profile specifies a scheduled time at which the switch is to occur, and this scheduled time can correspond to a start time for a real-world event that is known to cause an increase or decrease in traffic.