False positives for drive utilization

[stormonts]

We have a threshold setup that uses "Host MIB - Hard Drive Space" to alert when the "hdd_used" data source exceeds 70% of the "hdd_total" datasource. We will often receive alerts with a complete insane value such as "Used Space - / [hdd_used] [hdd_used] went above threshold of 70 with 1790.035". Looking at the Cacti logs doesn't show anything related to the host that threw the alert that would seem to correlate with the weird message. We are using thold .5 with Cacti 0.8.8f.