Getting stats for a VM seems to be different depending on the stat

While getting some stats for a vm the behavior seems to be different depending on the stat.
Is there any documentation for specific stats?

Sample Call:

metrics: controller_avg_read_io_size_kbytes,controller_avg_io_latency_usecs

startTimeInUsecs: 1631290740000000

intervalInSecs: 3600

Output (see below):
stat controller_avg_io_latency_usecs - give the expected output of a value for each interval since the input start time

but for stat controller_avg_read_io_size_kbytes - A single value is returned. What is the meaning of single value? Is it just the read at that point in time? Or is any calculation done behind the scenes?
 

{
  "statsSpecificResponses": [
    {
      "successful": true,
      "message": null,
      "startTimeInUsecs": 1632404100000000,
      "intervalInSecs": 3600,
      "metric": "controller_avg_read_io_size_kbytes",
      "values": [
        127
      ]
    },
    {
      "successful": true,
      "message": null,
      "startTimeInUsecs": 1631290740000000,
      "intervalInSecs": 3600,
      "metric": "controller_avg_io_latency_usecs",
      "values": [
        3587,
        3476,
        3514,
        3665,
        3700,
        3634,
        3688,
        3860 
       ]
    }
  ]
}