This article explores how to use INFINI Gateway's regex replacement capabilities to fix problematic query parameters. The techniques discussed here also apply to Opensaerch and Easysearch environments.

In a previous article, we covered the request_body_json_set processor for modifying query parameters. This piece will focus on the request_body_regex_replace processor, which offers a different approach to parameter adjustment.

Consider the following two search requests. If the goal is to modify only the second query's size parameter to 10, how would you approach this?

# Request A
GET test/_search
{
  "size": 50,
  "query": {
    "match": {
      "status": "active"
    }
  }
}

# Request B
GET test/_search
{
  "size": 5000,
  "query": {
    "match": {
      "status": "pending"
    }
  }
}

Using request_body_json_set would require conditional logic to target specific requests. The request_body_regex_replace processor, however, allows direct pattern matching and replacement for the specific data you need to modify.

request_body_regex_replace

The request_body_regex_replace filter in INFINI Gateway uses regular expressions to replace string content within the request body. For the queries above, you can add the request_body_regex_replace section to INFINI Gateway's default configuration as shown below.

flow:
  - name: default_flow
    filter:
      - request_body_regex_replace:
          pattern: '"size": 5000'
          to: '"size": 10'
      - elasticsearch:
          elasticsearch: prod
          max_connection_per_node: 1000

This configuration scans incoming search requests for the string "size": 5000 and replaces it with "size": 10.

Modification Results

After applying the request_body_regex_replace processor, the affected query appears as follows.

The modification was successful. As another practical example, modifying application query conditions typically requires code changes and can be time-consuming. The techniques shown here enable rapid query condition adjustments without touching application code.

Consider an application sending this search request:

GET products/_search
{
  "query": {
    "bool": {
      "filter": [
        {"term": {"type": "electronics" }}
      ],
      "must": [
        {"term": {"manufacturer": {"value": "acme" }}}
      ]
    }
  }
}

The query intends to retrieve electronics from manufacturer acme, but the type field incorrectly specifies "electronics" when it should be "computing". You can use request_body_regex_replace to correct the type field with this configuration:

flow:
  - name: default_flow
    filter:
      - request_body_regex_replace:
          pattern: '"type": "electronics"'
          to: '"type": "computing"'
      - elasticsearch:
          elasticsearch: prod
          max_connection_per_node: 1000

The configuration searches for "type": "electronics" in the request body and replaces it with "type": "computing".

As you can see, the request_body_regex_replace processor provides great flexibility. Evaluate your specific requirements to choose between request_body_regex_replace and request_body_json_set processors.