Repository navigation
NETWatcher 5.1.5
NETWatcher 5.1.5
This update makes outage detection faster and latency values more accurate. All targets are now checked at the same time, so a connection loss is confirmed within a single timeout instead of waiting for every failing target in turn. Ping latency is read from the ping result itself, which removes the delay of starting the ping process from the measured value.
The interface now receives only new chart points on each refresh instead of the full history, which lowers CPU use with long chart ranges. Monitoring cycles keep a steady interval, and each cycle's measurements are written to the log in one step.