Kibana date filter Elasticsearch error "failed to parse date field [2023-01-01T00:00:00.000+02:00] with format [dd/MM/yyyy HH:mm:ss Z||epoch_millis]

I have a date field in my Elasticsearch index mapped like:

      "updated_at": {
        "format" : "dd/MM/yyyy HH:mm:ss Z||epoch_millis",
        "type": "date"
      },

The value in the docs _source are epoch millis like "created_at": 1676689641549

If I use the Kibana 7.10 query interface to add a filter like this

This returns the error

> "failed to parse date field [2023-01-01T00:00:00.000+01:00] with format [dd/MM/yyyy HH:mm:ss Z||epoch_millis]"

Is there a setting in Kibana 7.10 or in my Elasticsearch mapping that makes the filter UI and the index query play nicely together?

NOTES:

• If I enter the values in milliseconds, the filer works correctly

• This is related to, but not solved by, this previous question /63352320/elasticsearch-parse-date-field-in-different-format

UPDATE 4. Jul 2023: The query generated by Kibana looks like this:

"query": {
    "bool": {
      "must": [],
      "filter": [
        {
          "match_all": {}
        },
        {
          "range": {
            "created_at": {
              "gte": "2022-01-01T00:00:00.000+01:00",
              "lt": "2023-01-01T00:00:00.000+01:00"
            }
          }
        }
      ],
      "should": [],
      "must_not": []
    }
  },

This just worked when I updated the mappings and reindexed with:

          "created_at": {
            "type": "date",
            "format": "date_time||epoch_millis"
          },

Thanks @val for the suggestion!