monitoring a connectivity status of one or more connectivity sources, the one or more connectivity sources comprising at least a first connectivity source and a second connectivity source;

selecting one of one or more available connectivity sources for use for online communications;

connecting a user's computer to a remote computing system via the selected available connectivity source;

monitoring whether the connection to the remote computing system via the selected connectivity source has failed;

if the connection is detected as failed, then scheduling a poll on a background software thread;

if the poll fails, then generating a notification that the connection to the remote computing system via the selected connectivity source is disconnected, wherein the selected connectivity source comprises the first connectivity source;

attempting reconnection to the remote computing system;

if the selected connectivity source is lost, determining whether the second connectivity source is available;

if the second connectivity source is available, automatically connecting the user's computer to the remote computing system via the second connectivity source without user action;

reducing an interval at which reconnection to the remote computing system is allowed from a first interval to a second interval after not being able to connect to the remote computing system using the first connectivity source for a given time period so that a connection attempt may be made to the remote computing system via the second connectivity source after the reduced interval;

resetting the time period upon detecting a network change, the network change comprising a hardware change at the user's computer;

switching back to the first interval upon detecting the network change; and

marking the one or more connectivity sources as inoperable while the user's computer is shutting down to prevent subsequent online communication events from adding to shutdown delays.