Discard stale information from DiskIO and NetworkIO meters

This ensures the initial read of the data is not misinterpreted as arbitrarily large values.
Also this forces the maximum update interval between two subsequent reads to within 30s.

Fixes #860

1 files changed, 6 insertions, 0 deletions

diff --git a/DiskIOMeter.c b/DiskIOMeter.c
index c5f14768..11fb791e 100644
--- a/DiskIOMeter.c
+++ b/DiskIOMeter.c
@@ -81,6 +81,12 @@ static void DiskIOMeter_updateValues(Meter* this) {
       cached_msTimeSpend_total = data.totalMsTimeSpend;
    }
 
+   if (passedTimeInMs > 30000) {
+      // Triggers for the first initialization and
+      // when there was a long time we did not collect updates
+      hasData = false;
+   }
+
    this->values[0] = cached_utilisation_diff;
    this->total = MAXIMUM(this->values[0], 100.0); /* fix total after (initial) spike */