Should I use Connection Pooling in AWS Lambda?

Question

I recently switched over an API from an EC2 instance to API Gateway with lambda. I noticed that my CPU usage on the database then shot up quite rapidly.

Below is an image of the active connections on the database, the red arrow is when I switched over to Lambda from an EC2. There where no new customers added so the actual throughput of data remained the same.

I am using the same code that I am using as when I was hosting on the Ec2 instance. This code uses a connection pool to connect to the database which I am not sure is the best idea. Basically I call the query method when I want to query the database in my main script.

A connection is retrieved from a connection pool for the lambda function to run the query and then released at the end of the query.

Is this a good approach or should one connection be constantly persisted throughout the execution of the lambda function? The CPU usage on a T3 medium is nearly maxed out since I moved to Lambda and I think this is the problem.

// This file is used to get a connection from the mysql connection pool 
// The pool is contained in this file and the function to get a connection is exported 

// Imports //
const mysql     = require('mysql')

// Define the connection pool //
const pool = mysql.createPool({
    host:       process.env.DBHOST,
    user:       process.env.DBUSER,
    password:   process.env.DBPASSWORD,
    port:       process.env.DBPORT,
    database:   process.env.DBNAME
})

/// getConnection ///
// Call this function whereever you need a connection
// It gets a connection from the connection pool 
function getConnection(){
    return new Promise((resolve,reject) => {
    pool.getConnection((err,con) => {
        if(err){
            reject(`Connection Failed\n${err}`);
        }else{
            console.info(`Successful connection, id ${con.threadId}`)
            resolve(con)
        }
    })
})
}

/// query ///
// This function makes a query on a database 
// It accepts and sql statement and a list of args if used in the statemens
// Inputs:
//  sql  - The sql statement to execute on the database 
//  args - The arguments to use in the sql statament  
// Returns:
//  resolve - The result of the sql insert statemen 
//  reject  - An error
async function query(sql,args){
    return new Promise(async(resolve,reject) => {
        // try/catch is to handle the await for the connection 
        try{
            let con = await getConnection() // If there is an error here it will be caught
            // No error this query will execute 
            con.query(sql,args, function(err,result) {
                // If error returned from query will be caught here  
                if (err){
                    reject(err);// Reject the promise 
                }else{
                    resolve(result);// Resolve the promise 
                }
                // Always release
                con.release() 
            })
        }catch(error){
            // Error in getting connection reject the error 
            reject(error)
        }
    })    
}

module.exports = {query}