DomInfoUpdateTask polling can take longer than DOM_INFO_UPDATE_PERIOD_SECS due to 1 sec wait after every iteration

[aditya-nexthop]

DomInfoUpdateTask polling loop is structured as follows:

            for physical_port, logical_ports in self.port_mapping.physical_to_logical.items():
                # Process pending link change events and update diagnostic
                # information in the database. Ensures timely handling of link
                # change events and avoids duplicate updates in case of breakout ports.
                port_change_observer.handle_port_update_event() <<=============
                # Process each port in the pending link change set based on the
                # corresponding time to update the DB after the link change.
                for link_changed_port in list(self.link_change_affected_ports.keys()):
                    if self.task_stopping_event.is_set():
                        self.log_notice("Stop event generated during DOM link change event processing")
                        break
                    if self.link_change_affected_ports[link_changed_port] <= datetime.datetime.now():
                        self.log_notice(f"Updating port db diagnostics post link change for port {link_changed_port}")
                        self.update_port_db_diagnostics_on_link_change(link_changed_port)
                        del self.link_change_affected_ports[link_changed_port]

                if self.task_stopping_event.is_set():
                    self.log_notice("Stop event generated during DOM monitoring loop")
                    break

                if not is_periodic_db_update_needed:
                    # If periodic db update is not needed, skip the rest of the loop
                    continue

                # Get the first logical port name since it corresponds to the first subport
                # of the breakout group
                logical_port_name = logical_ports[0]
...

port_change_observer.handle_port_update_event() call goes to xcvrd/xcvrd_utilities/port_event_helper.py:

    def handle_port_update_event(self):
        """
        Select PORT update events, notify the observers upon a port update in CONFIG_DB
        or a XCVR insertion/removal in STATE_DB

        Returns:
            bool: True if there's at least one update event; False if there's no update event.
        """
        has_event = False
        if not self.stop_event.is_set():
            (state, _) = self.sel.select(SELECT_TIMEOUT_MSECS)
            if state == swsscommon.Select.TIMEOUT:
                return has_event
            if state != swsscommon.Select.OBJECT:
                self.logger.log_warning('sel.select() did not return swsscommon.Select.OBJECT')
                return has_event
...

Here, SELECT_TIMEOUT_MSECS = 1000

So when there are no port change events, we spend 1 sec waiting for an event in handle_port_update_event.
On systems with a large number of transceivers or with many VDM enabled transceivers (where polling takes longer), we are unable to poll information from all transceivers in DOM_INFO_UPDATE_PERIOD_SECS (currently 60 secs)

This leads to flags and DOM information updating slower than 60 seconds.