ProGlove Documentation

Retrieve the number of scans for a use case in time frames of one hour grouped by the used scanning device:

{
    "range_start": "2021-02-20T00:00:00+0000",
    "range_end": "2021-02-22T23:59:59+0000",
    "resolution": "1h",
    "select": [
      {"name": "measure_value::varchar", "aggregate": "count", "alias": "scans"},
      {"name": "gateway_l3_id"}
    ],
    "where": [
      {"name": "measure_name", "operator": "=", "value": "scan_code"},
      {"name": "gateway_l2_id", "operator": "=", "value": "b2a931"}
    ],
    "groupby": ["gateway_l3_id"]
}

The sections below describe the existing constraints:

The time-series data is always selected within the bounds of a time frame. This is specified by two inclusive points in time, range_start and range_end, as the boundaries. Only data within the interval of these points in time is considered for the query and its result.

The resulting time-series data is grouped into equidistant time slots. By picking a suitable resolution, the granularity of this grouping can be adjusted to fit the time range and observation requirements.

To group the data into time windows the following resolution values are available:

To select measurements:

E.g. this selects the scan_duration measure:

{
    "select": [
      {"name": "measure_value::double", "aggregate": "sum"},
    ],
    "where": [
      {"name": "measure_name", "aggregate": "=", "value": "scan_duration"},
    ]
}

To aggregate data, you can use one of these functions:

These aggregation functions are applied to all selected data with a single group and time resolution.

The resulting data can be grouped by:

Dividing the data for analysis only into groups of time windows or attribute values is often not enough as it does not allow you to zoom in on specific data. Instead, before grouping and aggregating the data, a filtering pass is necessary.

E.g. consider the retrieved data should be that of a specific process or a set of devices. To achieve this, a filtering clause can be used to exclude any data not matching the criteria:

At the moment data can be matched only by comparing for equality of a string value. To match data attributes these string-valued operators can be used:

The order of the returned data can be selected with an ordering clause: One or more attributes can be selected to order the data by. In addition, the data can be either returned in an ascending (lowest values first) or descending direction.

An important use of data ordering is selecting the tail or head of some data: E.g. the top 5 of some measurement or the lowest 10 of some datum can be implemented by ordering such that the first returned rows are the wanted ones and selecting a limit to get exactly the wanted rows.

{
    "range_start": "2021-02-20T00:00:00+0000",
    "range_end": "2021-02-22T23:59:59+0000",
    "select": [
      {"name": "measure_value::double", "aggregate": "sum", "alias": "duration"},
      {"name": "gateway_l3_id"}
    ],
    "where": [
      {"name": "measure_name", "aggregate": "=", "value": "scan_duration"},
    ],
    "groupby": ["gateway_l3_id"],
    "orderby": ["duration"],
}