Time series data points

When creating the time series attribute definition, you can specify the data type that it can hold with the bandFactor parameter, which determines the precision (number of decimal places) of stored value.

The possible bandFactor values are 1, 0.1, 0.01, 0.001, and 0.0001.

The bandFactor is applied to the data value by dividing data value by bandFactor. The resulting number is then rounded to a whole number. This whole number is called a normalized value, and is stored in the database.

Any invalid time series data value (string or value beyond the supported range) is stored as a data hole in the Analyzer detail view database.

Time series data interval

The start time must be the multiple of the data interval. If not, then the next highest multiple of the data interval is considered as the start time.

For example, the data interval is 5 minutes and the start time is 20150322_051700, which is not a multiple of 5 minutes. In this case, the Analyzer detail view database shifts the start time to the next multiple of 5 minutes, which is 20150322_052000.

Data for the same resource and the same time series counter can be added with a different data interval in the same time range.

For example, insert data for different intervals (1 and 60) for the same resource, counter, and time:

{
  "signature":"vm#cumulus-labvm1",
  "name":"cumulus-lab-vm1",
  "cpuUsage":[
    {
      "from":"20141217_020000",
      "interval":1,
      "data":[
        1, 2, 3, 0
      ]
    },
    {
      "from":"20141217_020000",
      "interval":60,
      "data":[
        23.5, 33, 67.5, 4
      ]
    }
  ]
}