Introduction

Metrics are aggregated values on the specified dimensions.
Two types of metrics exist: basic metric fields and custom aggregations.

Metric fields

A metric field is a plain field on which the aggregation operation is pre-defined. For example:

{
  ...,
  "query": {
    ...,
    "metrics": [
      "crawl.20210102.count_urls_crawl",
      "search_console.period_0.count_clicks"
    ],
    ...
  }
}

• crawl.20210102.count_urls_crawl will count the number of different URLs that were crawled
• search_console.period_0.count_clicks will sum the number of clicks that occurred on the specified dimensions

Here the aggregations are pre-defined, since you can only count the URLs and sum the clicks respectively.

Custom aggregations

For some fields, we require more flexibility and one can choose the aggregation that should be applied. This is done by applying a function to the field:

{
  ...,
  "query": {
    ...,
    "metrics": [
      {
        "function": "avg",
        "args": ["crawl.20210102.depth"]
      }
    ],
    ...
  }
}

which will compute the average URL depth on each specified dimension.

Supported custom aggregation functions:

• sum
• avg
• min
• max
• weighted_avg (takes two input arguments)
• count
• count_distinct
• count_distinct_approx
• count_true
• count_false
• count_null
• count_eq (takes two input arguments)
• count_gt (takes two input arguments)
• count_gte (takes two input arguments)
• count_lt (takes two input arguments)
• count_lte (takes two input arguments)