http-aws-es compatible with new Elasticsearch client.

aws_es_connector.js

/* requires AWS creds to be updated.
 * if they aren't, update using AWS.config.update() method before instatiing the client. 
 * 
 * import this module where you instantiate the client, and simply pass this module as the connection class. 
 *
 * eg:
 * const client = new Client({
 *   node,
 *   Connection: AwsConnector
 * });
 */
 
import AWS from 'aws-sdk';
import { Connection } from '@elastic/elasticsearch';

class AwsConnector extends Connection {
  async request(params, callback) {
    try {
      const creds = await this.getAWSCredentials();
      const req = this.createRequest(params);

      const { request: signedRequest } = this.signRequest(req, creds);
      super.request(signedRequest, callback);
    } catch (error) {
      throw error;
    }
  }

  createRequest(params) {
    const endpoint = new AWS.Endpoint(this.url.href);
    let req = new AWS.HttpRequest(endpoint);

    Object.assign(req, params);
    req.region = AWS.config.region;

    if (!req.headers) {
      req.headers = {};
    }

    let body = params.body;

    if (body) {
      let contentLength = Buffer.isBuffer(body)
        ? body.length
        : Buffer.byteLength(body);
      req.headers['Content-Length'] = contentLength;
      req.body = body;
    }
    req.headers['Host'] = endpoint.host;

    return req;
  }

  getAWSCredentials() {
    return new Promise((resolve, reject) => {
      AWS.config.getCredentials((err, creds) => {
        if (err) {
          if (err && err.message) {
            err.message = `AWS Credentials error: ${e.message}`;
          }

          reject(err);
        }

        resolve(creds);
      });
    });
  }

  signRequest(request, creds) {
    const signer = new AWS.Signers.V4(request, 'es');
    signer.addAuthorization(creds, new Date());
    return signer;
  }
}

export { AwsConnector };

es_client.js

import { AwsConnector } from './aws_es_connector';
import { Client } from '@elastic/elasticsearch';
import AWS from 'aws-sdk';

// load aws keys and region, ideally using env variables.
let accessKey = '****';
let secretKey = '****'; 
let region = '**'; //eg: us-east-1
let node = '***'; //node url

AWS.config.update({
  credentials: new AWS.Credentials(accessKey, secretKey),
  region
});

const client = new Client({
  node,
  Connection: AwsConnector
});

//use this client to talk to AWS managed ES service.
export default client; 

Figured it out.
AWSHTTPRequest uses path to send query parameters and there is no key as "querystring" here in docs.
So a small change in aws_es_connector.js solves the problem.
In createRequest() before sending the req object
req.path+='/?'+req.querystring; delete req.querystring;
Now all parameters will work fine.

cool .thanks saved my day

Thank you @parthdesai93! Was very helpful

Hey. I find I spot an issue.
Look at this:
`

const x = 'ߤߤߤߤ';
x.length
4
w = Buffer.from(x)

> w.byteLength 8 `

Content-Length is provided in bytes. So the header is going to have invalid value - 4 bytes instead of 8.