Currently, the start date provided to sreport is just the earliest start date across a group's active allocations. This is not necessarily accurate as the usage value being compared against the limits is the usage since the last raw usage reset.
This PR take the rawusage reset date into consideration, by returning the max of the earliest start date found and the most recent raw usage reset date. We want the max because start dates found prior to the raw usage reset date are what's creating the possibility of > 100% usage in the crc-usage output.
I tested this works properly on the cluster, and that it does not break crc-sus.
Currently, the start date provided to sreport is just the earliest start date across a group's active allocations. This is not necessarily accurate as the usage value being compared against the limits is the usage since the last raw usage reset.
This PR take the rawusage reset date into consideration, by returning the max of the earliest start date found and the most recent raw usage reset date. We want the max because start dates found prior to the raw usage reset date are what's creating the possibility of > 100% usage in the crc-usage output.
I tested this works properly on the cluster, and that it does not break
crc-sus
.Before:
After: