Average number of info logs in a particular time period

So this example query would return the documents between 6 and 9 bucketed by day:

GET logstash-0/_search
{
  "aggs": {
    "by_day": {
      "date_histogram": {
        "field": "@timestamp",
        "interval": "day",
        "format": "yyyy-MM-dd"
      },
      "aggs": {
        "between6and9": {
          "filter": {
            "script": {
              "script": {
                "source": "int hour = doc['@timestamp'].value.hourOfDay; hour > 17 && hour < 22",
                "lang": "painless"
              }
            }
          }
        }
      }
    }
  }
}

You could save it as a custom query and then create a visualization against it.