I believe there is an issue with how ping requests are processed.

According to the source visit_last_action_time holds the visit´s end time:

/**
 * This dimension holds the best guess for a visit's end time. It is set the last action
 * time for each visit. `ping=1` requests can be sent to update the dimension value so
 * it can be a more accurate guess of the time the visitor spent on the site.
 *
 * Note: though it is named 'visit last action time' it actually refers to the visit's last action's
 * end time.
 */

https://github.com/matomo-org/matomo/blob/6b4cd1676d524641167ba32776a999e3e37a6a2a/plugins/CoreHome/Columns/VisitLastActionMinute.php

Looking at how the PingRequestProcessor (https://github.com/matomo-org/matomo/blob/3.x-dev/plugins/Heartbeat/Tracker/PingRequestProcessor.php) handles hearbeats, it seems it does not update log_visit.visit_last_action_time, it only updates log_visit.visit_total_time.

log_visit.visit_last_action_time is only updated if an action is sent. I think the issue is that PingRequestProcessor sends an empty Actions:
https://github.com/matomo-org/matomo/blob/2.16.0-b6/plugins/Heartbeat/Tracker/PingRequestProcessor.php#L24-L29

When viewing a users visits, the "x actions in " is correct, as it uses log_visit.visit_first_action_time + log_visit. visit_total_time.

Example of this:

For us, this becomes an issue an issue when using CustomReports and Server Time - Hour/Minute (End Of Visit), as it using log_visit.visit_last_action_time (https://github.com/matomo-org/matomo/blob/3.x-dev/plugins/CoreHome/Columns/VisitLastActionMinute.php#L32) to determine the visis end time. This occurs since ping requests does not update log_visit.visit_last_action_time. With the current code, it is not possible to determine when a visit ended for SPA apps, since it seems only an action can that trigger an update of log_visit.visit_last_action_time.

I'm not sure if this is a bug in the PingRequestProcessor, or an issue with how VisitLastActionMinute is expected to compute the visits end time.

Issues i think are relevant:
#9626 #9748