WIP: benchmarks (#745)

* Updated dependencies

* Updated .gitignore

* WIP: macro and micro benchmarks

* Updated benchmark suite

* Use the same suite for both macro and micro benchmarks

* WIP: benchmark report

* Updated benchmark suite

* Updated docker scripts

* Updated benchmark suite

* Updated scripts

* Updated benchmark suite

* Added split2

test/benchmarks/micro/basic.bench.js

@@ -0,0 +1,98 @@
+'use strict'
+
+const { bench } = require('../suite')({
+  report: {
+    url: process.env.ES_RESULT_CLUSTER_URL,
+    username: process.env.ES_RESULT_CLUSTER_USERNAME,
+    password: process.env.ES_RESULT_CLUSTER_PASSWORD
+  }
+})
+const { Client } = require('../../../index')
+const { connection } = require('../../utils')
+
+bench('Initialization', { warmup: 5, measure: 10, iterations: 1000 }, async b => {
+  b.start()
+  for (var i = 0; i < b.iterations; i++) {
+    const client = new Client({ // eslint-disable-line
+      node: 'http://localhost:9200'
+    })
+  }
+  b.end()
+})
+
+bench('Call api with lazy loading', { warmup: 5, measure: 10 }, async b => {
+  const client = new Client({
+    node: 'http://localhost:9200',
+    Connection: connection.MockConnection
+  })
+
+  b.start()
+  await client.info()
+  b.end()
+})
+
+bench('Call api without lazy loading', { warmup: 5, measure: 10 }, async b => {
+  const client = new Client({
+    node: 'http://localhost:9200',
+    Connection: connection.MockConnection
+  })
+
+  await client.info()
+  b.start()
+  await client.info()
+  b.end()
+})
+
+bench('Basic get', { warmup: 5, measure: 10, iterations: 1000 }, async b => {
+  const client = new Client({
+    node: 'http://localhost:9200',
+    Connection: connection.MockConnection
+  })
+
+  // we run the method twice to skip the lazy loading overhead
+  await client.search({
+    index: 'test',
+    type: 'doc',
+    q: 'foo:bar'
+  })
+  b.start()
+  for (var i = 0; i < b.iterations; i++) {
+    await client.search({
+      index: 'test',
+      type: 'doc',
+      q: 'foo:bar'
+    })
+  }
+  b.end()
+})
+
+bench('Basic post', { warmup: 5, measure: 10, iterations: 1000 }, async b => {
+  const client = new Client({
+    node: 'http://localhost:9200',
+    Connection: connection.MockConnection
+  })
+
+  // we run the method twice to skip the lazy loading overhead
+  await client.search({
+    index: 'test',
+    type: 'doc',
+    body: {
+      query: {
+        match: { foo: 'bar' }
+      }
+    }
+  })
+  b.start()
+  for (var i = 0; i < b.iterations; i++) {
+    await client.search({
+      index: 'test',
+      type: 'doc',
+      body: {
+        query: {
+          match: { foo: 'bar' }
+        }
+      }
+    })
+  }
+  b.end()
+})