Adding {{period_start}} or {{period_end}} give `all shards failed` error?

This is a bit of a weird one, if I pass through an alert with aggregations during the Define Monitor section of creating a monitor, I get the following behaviour:

      "must": [
        {
          "range": {
            "@timestamp": {
              "format": "strict_date_optional_time",
              "gte": "2020-03-18T23:54:53.351Z",
              "lte": "2020-03-19T05:54:53.351Z"
            }
          }
        }

It works fine without issue and returns results. If I do the same query, with the interval set as a 6 hour period, with the following:

      "must": [
        {
          "range": {
            "@timestamp": {
              "format": "strict_date_optional_time",
              "gte": "{{period_start}}",
              "lte": "{{period_end}}"
            }
          }
        }
      ],

I get a response back from the API stating:

{"ok":true,"resp":{"monitor_name":"TEMP_MONITOR","period_start":1584576246230,"period_end":1584597846230,"error":null,"input_results":{"results":[],"error":"all shards failed"},"trigger_results":{}}}

Anyone else ever experienced this?

Looks like the peiord_start and period_end dates are crazy out of whack for some reason