ml_job_raw_data_summary_watch.json

#watch that counts number of anomalies and number of docs in an index
POST _xpack/watcher/watch/_execute
{
  "watch": {
    "trigger": {
      "schedule": {
        "interval": "5m"
      }
    },
    "metadata": {
      "watch_timespan" : "7d",    
      "job_name" : "a_kibana_sample_data_flights",     
      "job_min_record_score": 0
     },
    "input": {
      "chain": {
        "inputs": [
          {
            "ml_job": {
              "search": {
                "request": {
                  "indices": [
                    ".ml-anomalies-*"
                  ],
                  "body": {
                    "size":0,
                    "query": {
                      "bool": {
                        "filter": [
                          { "range": { "timestamp": {"gte": "now-{{ctx.metadata.watch_timespan}}"}}},
                          { "term":  {"result_type": "record"}},
                          { "term":  {"job_id": "{{ctx.metadata.job_name}}"}},
                          { "range": {"record_score": {"gte": "{{ctx.metadata.job_min_record_score}}"}}}
                        ]
                      }
                    }
                  }
                }
              }
            }
          },
          {
            "raw_data": {
              "search": {
                "request": {
                  "indices": [
                    "kibana_sample_data_flights"
                  ],
                  "body": {
                    "size":0,
                    "query": {
                      "bool": {
                        "filter": [
                          { "range": { "timestamp": {"gte": "now-{{ctx.metadata.watch_timespan}}"}}}
                        ]
                      }
                    }
                  }
                }
              }
            }
          }
        ]
      }
    },
  "condition" : {
    "script" : {
      "source" : "return ctx.payload.ml_job.hits.total > 0 && ctx.payload.raw_data.hits.total > 0"
    }
  },
    "actions": {
      "log": {
        "transform": {
          "script": "return ['num_anomalies': ctx.payload.ml_job.hits.total,'num_docs':ctx.payload.raw_data.hits.total, '@timestamp': ctx.execution_time ]"
        },
        "logging": {
          "text": "num_anomalies={{ctx.payload.num_anomalies}},num_docs={{ctx.payload.num_docs}} seen when executed at {{ctx.payload.@timestamp}} "
        }
      },
     "index_payload": {
      "transform": {
        "script": "return ['num_anomalies': ctx.payload.ml_job.hits.total,'num_docs':ctx.payload.raw_data.hits.total, '@timestamp': ctx.execution_time ]"
      },
      "index": {
        "index": "my-summary",
        "doc_type": "_doc"
      }
    }
    }
  }
}

GET my-summary/_search

returns:

{
  "took" : 0,
  "timed_out" : false,
  "_shards" : {
    "total" : 1,
    "successful" : 1,
    "skipped" : 0,
    "failed" : 0
  },
  "hits" : {
    "total" : {
      "value" : 1,
      "relation" : "eq"
    },
    "max_score" : 1.0,
    "hits" : [
      {
        "_index" : "my-summary",
        "_type" : "_doc",
        "_id" : "USPXBGwBvJSK0DlwR-0I",
        "_score" : 1.0,
        "_source" : {
          "num_docs" : 7539,
          "num_anomalies" : 1,
          "@timestamp" : "2019-07-18T11:28:32.875Z"
        }
      }
    ]
  }
}

Obviously, one would make watch_timespan equal to the trigger interval for on-going execution