How does the flow works in elasticsearch queries?

I have written a query which has couple of condition as shown below.

     GET /agreement/_search
        {
          "query": {
            "bool": {
              "must": [
                {
                  "multi_match": {
                    "query": "T-0668",
                    "fields": [
                      "agreecondition.agreementId",
                      "agreecondition.conditionContractId"
                    ]
                  }
                },
    ,
            {
              "range": {
                "agreecondition.validFrom": {
                  "gte": "02/18/2019"
                }
              }
            },
            {
              "range": {
                "agreecondition.validTo": {
                  "lte": "03/07/2019"
                }
              }
            }
              ],
              "filter": [
                {
                  "terms": {
                    "agreecondition.promotionId.keyword": [
                    "x",
"y"
                    ]
                  }
                }
              ]
            }
          }
        }

My question is how the flow works?

Ex: Does the ES first gets the results for the must condition's multi-match and on the output of the multi-match, does the range condition applies? followed by filter(on top of the output of the range condition)?

I just wanted to get a clarity on this, if my assumption is wrong, then i need to re-write the query.

Solution

You can check elasticsearch official blog on query execution order to understand this in details but you might just not get all the details you are looking for, due to limitation elastic put as mentioned at the end of the blog:

Q: How can I check which query/filter got executed first?

A: We don't really expose this information, which is very internal. However if you

check the output of the profile API, you can count how many times nextDoc/advance have been called on the one hand, and matches on the other hand. Query nodes that have the higher counts have been run first.

Note: Profile API will be very handful for you as suggested in the blog as well.