ChameleonKVMCloudTraces.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Chameleon KVM@TACC Cloud Traces"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Chameleon provides an installation of OpenStack Kilo using the KVM virtualization technology at the [KVM@TACC][1] site. For more information about the Chameleon KVM site, please read our [user documentation][2]. \n",
    "\n",
    "Chameleon Team has been developed a [trace data structure][3] based on data from OpenStack Nova. In addition, a [software tool][4] has been implemented and used to extract the appropriate data. The Chameleon [KVM@TACC][1] cloud traces are then published on [Science Clouds][4].\n",
    "\n",
    "In this Jupyter Notebook, we show simple examples of how to download Chameleon KVM cloud traces and analyze the data. \n",
    "\n",
    "[1]: https://openstack.tacc.chameleoncloud.org\n",
    "[2]: https://chameleoncloud.readthedocs.io/en/latest/technical/kvm.html\n",
    "[3]:https://scienceclouds.org/cloud-traces/cloud-trace-format/\n",
    "[4]: https://github.com/ChameleonCloud/starcompactor\n",
    "[5]: http://press3.mcs.anl.gov/scienceclouds/cloud-traces/"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Download Cloud Traces\n",
    "\n",
    "Download Chameleon KVM@TACC cloud traces from [Science Clouds][1]. \n",
    "\n",
    "[1]: http://press3.mcs.anl.gov/scienceclouds/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading and reading cloud traces completed successfully!\n"
     ]
    }
   ],
   "source": [
    "import datetime\n",
    "import pandas as pd \n",
    "\n",
    "from io import BytesIO\n",
    "from zipfile import ZipFile\n",
    "from urllib.request import urlopen\n",
    "\n",
    "resp = urlopen(\"http://scienceclouds.org/files/2018/12/Chameleon_KVM_cloud_trace.zip\")\n",
    "zipfile = ZipFile(BytesIO(resp.read()))\n",
    "# zip contains two file - instance events and machine events\n",
    "machine_events_df = pd.read_csv(zipfile.open('machine_events.csv'), parse_dates=['EVENT_TIME'])\n",
    "instance_events_df = pd.read_csv(zipfile.open('instance_events.csv'), parse_dates=['START_TIME', 'FINISH_TIME'])\n",
    "\n",
    "print(\"Downloading and reading cloud traces completed successfully!\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>EVENT_TIME</th>\n",
       "      <th>EVENT_TIME_SEC</th>\n",
       "      <th>HOST_NAME (PHYSICAL)</th>\n",
       "      <th>RACK</th>\n",
       "      <th>EVENT</th>\n",
       "      <th>VCPU_CAPACITY</th>\n",
       "      <th>MEMORY_CAPACITY_MB</th>\n",
       "      <th>DISK_CAPACITY_GB</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2015-11-05 05:05:01</td>\n",
       "      <td>4251901.0</td>\n",
       "      <td>d99a3d1c0af903877c7374a1845f8859</td>\n",
       "      <td>1</td>\n",
       "      <td>ENABLE</td>\n",
       "      <td>48</td>\n",
       "      <td>128536</td>\n",
       "      <td>156336</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2017-07-31 00:12:45</td>\n",
       "      <td>59011965.0</td>\n",
       "      <td>803963824739adc8b8c6d3d4d03b3b0a</td>\n",
       "      <td>2</td>\n",
       "      <td>UPDATE</td>\n",
       "      <td>48</td>\n",
       "      <td>128536</td>\n",
       "      <td>156336</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2015-10-09 21:32:34</td>\n",
       "      <td>1978354.0</td>\n",
       "      <td>1d659223dcc41970623494c921d38198</td>\n",
       "      <td>2</td>\n",
       "      <td>CREATE</td>\n",
       "      <td>48</td>\n",
       "      <td>128536</td>\n",
       "      <td>156336</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2015-11-05 05:04:57</td>\n",
       "      <td>4251897.0</td>\n",
       "      <td>37d24a65ac407f6fe8ac883493e02512</td>\n",
       "      <td>2</td>\n",
       "      <td>ENABLE</td>\n",
       "      <td>48</td>\n",
       "      <td>128536</td>\n",
       "      <td>156336</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2015-11-04 05:04:53</td>\n",
       "      <td>4165493.0</td>\n",
       "      <td>8c8c2bc2c29645f3dfb27ea93f610154</td>\n",
       "      <td>1</td>\n",
       "      <td>DISABLE</td>\n",
       "      <td>48</td>\n",
       "      <td>128536</td>\n",
       "      <td>156336</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>2015-10-09 21:15:46</td>\n",
       "      <td>1977346.0</td>\n",
       "      <td>d114f0a789b6d3d6849840547df945b8</td>\n",
       "      <td>2</td>\n",
       "      <td>CREATE</td>\n",
       "      <td>48</td>\n",
       "      <td>128536</td>\n",
       "      <td>156336</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>2015-11-05 05:05:01</td>\n",
       "      <td>4251901.0</td>\n",
       "      <td>867affe3494a18886e0c1091ebbd8b8b</td>\n",
       "      <td>2</td>\n",
       "      <td>ENABLE</td>\n",
       "      <td>48</td>\n",
       "      <td>128536</td>\n",
       "      <td>156336</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>2017-04-26 04:04:53</td>\n",
       "      <td>50731493.0</td>\n",
       "      <td>11de6245d3e631ff22327793514aa832</td>\n",
       "      <td>0</td>\n",
       "      <td>DISABLE</td>\n",
       "      <td>40</td>\n",
       "      <td>64156</td>\n",
       "      <td>29689</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>2018-01-18 05:05:01</td>\n",
       "      <td>73803901.0</td>\n",
       "      <td>e711f2ed94b2fdb9b2b274af105cfcdc</td>\n",
       "      <td>1</td>\n",
       "      <td>UPDATE</td>\n",
       "      <td>48</td>\n",
       "      <td>128528</td>\n",
       "      <td>156336</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>2015-10-09 21:27:35</td>\n",
       "      <td>1978055.0</td>\n",
       "      <td>872b22df7a9304f65fa85e245ca02979</td>\n",
       "      <td>2</td>\n",
       "      <td>CREATE</td>\n",
       "      <td>48</td>\n",
       "      <td>128536</td>\n",
       "      <td>156336</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           EVENT_TIME  EVENT_TIME_SEC              HOST_NAME (PHYSICAL)  RACK  \\\n",
       "0 2015-11-05 05:05:01       4251901.0  d99a3d1c0af903877c7374a1845f8859     1   \n",
       "1 2017-07-31 00:12:45      59011965.0  803963824739adc8b8c6d3d4d03b3b0a     2   \n",
       "2 2015-10-09 21:32:34       1978354.0  1d659223dcc41970623494c921d38198     2   \n",
       "3 2015-11-05 05:04:57       4251897.0  37d24a65ac407f6fe8ac883493e02512     2   \n",
       "4 2015-11-04 05:04:53       4165493.0  8c8c2bc2c29645f3dfb27ea93f610154     1   \n",
       "5 2015-10-09 21:15:46       1977346.0  d114f0a789b6d3d6849840547df945b8     2   \n",
       "6 2015-11-05 05:05:01       4251901.0  867affe3494a18886e0c1091ebbd8b8b     2   \n",
       "7 2017-04-26 04:04:53      50731493.0  11de6245d3e631ff22327793514aa832     0   \n",
       "8 2018-01-18 05:05:01      73803901.0  e711f2ed94b2fdb9b2b274af105cfcdc     1   \n",
       "9 2015-10-09 21:27:35       1978055.0  872b22df7a9304f65fa85e245ca02979     2   \n",
       "\n",
       "     EVENT  VCPU_CAPACITY  MEMORY_CAPACITY_MB  DISK_CAPACITY_GB  \n",
       "0   ENABLE             48              128536            156336  \n",
       "1   UPDATE             48              128536            156336  \n",
       "2   CREATE             48              128536            156336  \n",
       "3   ENABLE             48              128536            156336  \n",
       "4  DISABLE             48              128536            156336  \n",
       "5   CREATE             48              128536            156336  \n",
       "6   ENABLE             48              128536            156336  \n",
       "7  DISABLE             40               64156             29689  \n",
       "8   UPDATE             48              128528            156336  \n",
       "9   CREATE             48              128536            156336  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# show machine events\n",
    "machine_events_df.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>INSTANCE_UUID</th>\n",
       "      <th>EVENT</th>\n",
       "      <th>START_TIME</th>\n",
       "      <th>START_SEC</th>\n",
       "      <th>FINISH_TIME</th>\n",
       "      <th>FINISH_SEC</th>\n",
       "      <th>EVENT_DURATION</th>\n",
       "      <th>RESULT</th>\n",
       "      <th>MEMORY_MB</th>\n",
       "      <th>DISK_GB</th>\n",
       "      <th>VCPUS</th>\n",
       "      <th>INSTANCE_NAME</th>\n",
       "      <th>USER_ID</th>\n",
       "      <th>PROJECT_ID</th>\n",
       "      <th>HOST_NAME (PHYSICAL)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>467af885e36bbb3a656fb8f83f8c91db</td>\n",
       "      <td>compute__do_build_and_run_instance</td>\n",
       "      <td>2015-09-17 18:24:56</td>\n",
       "      <td>66296.0</td>\n",
       "      <td>2015-09-17 18:25:06</td>\n",
       "      <td>66306.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>Success</td>\n",
       "      <td>512</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2c6824f29872b7ef0729973411589fc6</td>\n",
       "      <td>c358c86043c749283d1c498516e89b9a</td>\n",
       "      <td>54fdd12fff1505c8329c6fc50d4bcebc</td>\n",
       "      <td>40819f1d02e0ce5401227d282b5f58ad</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>84483962a31f776cf481fa56ff4babde</td>\n",
       "      <td>compute__do_build_and_run_instance</td>\n",
       "      <td>2015-09-17 20:25:55</td>\n",
       "      <td>73555.0</td>\n",
       "      <td>2015-09-17 20:25:58</td>\n",
       "      <td>73558.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>Success</td>\n",
       "      <td>512</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>fc5dac6c2f4d2a3223e4d298b3677740</td>\n",
       "      <td>c358c86043c749283d1c498516e89b9a</td>\n",
       "      <td>54fdd12fff1505c8329c6fc50d4bcebc</td>\n",
       "      <td>2412e43489814bc20cb5382e66e4d1f0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>84483962a31f776cf481fa56ff4babde</td>\n",
       "      <td>compute_terminate_instance</td>\n",
       "      <td>2015-09-17 20:26:19</td>\n",
       "      <td>73579.0</td>\n",
       "      <td>2015-09-17 20:26:19</td>\n",
       "      <td>73579.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>Success</td>\n",
       "      <td>512</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>fc5dac6c2f4d2a3223e4d298b3677740</td>\n",
       "      <td>c358c86043c749283d1c498516e89b9a</td>\n",
       "      <td>54fdd12fff1505c8329c6fc50d4bcebc</td>\n",
       "      <td>2412e43489814bc20cb5382e66e4d1f0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>974a749513018784b636dd1d6c2f435e</td>\n",
       "      <td>compute__do_build_and_run_instance</td>\n",
       "      <td>2015-09-17 20:26:42</td>\n",
       "      <td>73602.0</td>\n",
       "      <td>2015-09-17 20:28:00</td>\n",
       "      <td>73680.0</td>\n",
       "      <td>78.0</td>\n",
       "      <td>Success</td>\n",
       "      <td>4096</td>\n",
       "      <td>40</td>\n",
       "      <td>2</td>\n",
       "      <td>fc5dac6c2f4d2a3223e4d298b3677740</td>\n",
       "      <td>c358c86043c749283d1c498516e89b9a</td>\n",
       "      <td>54fdd12fff1505c8329c6fc50d4bcebc</td>\n",
       "      <td>2412e43489814bc20cb5382e66e4d1f0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>6ae938b5e39b74924f9a5a234d6fc8aa</td>\n",
       "      <td>compute__do_build_and_run_instance</td>\n",
       "      <td>2015-09-18 20:41:41</td>\n",
       "      <td>160901.0</td>\n",
       "      <td>2015-09-18 20:41:46</td>\n",
       "      <td>160906.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>Success</td>\n",
       "      <td>512</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>cdeeaab9e1a7158375537018afb215d9</td>\n",
       "      <td>38fb9a0d17479f503be2d51270061280</td>\n",
       "      <td>54fdd12fff1505c8329c6fc50d4bcebc</td>\n",
       "      <td>c8a66d52ecfce75831966c472449e4dc</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>6ae938b5e39b74924f9a5a234d6fc8aa</td>\n",
       "      <td>compute_terminate_instance</td>\n",
       "      <td>2015-09-18 20:53:36</td>\n",
       "      <td>161616.0</td>\n",
       "      <td>2015-09-18 20:53:37</td>\n",
       "      <td>161617.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Success</td>\n",
       "      <td>512</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>cdeeaab9e1a7158375537018afb215d9</td>\n",
       "      <td>38fb9a0d17479f503be2d51270061280</td>\n",
       "      <td>54fdd12fff1505c8329c6fc50d4bcebc</td>\n",
       "      <td>c8a66d52ecfce75831966c472449e4dc</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>334b1cfe61c72a30619663a4dc44da49</td>\n",
       "      <td>compute__do_build_and_run_instance</td>\n",
       "      <td>2015-09-18 20:53:58</td>\n",
       "      <td>161638.0</td>\n",
       "      <td>2015-09-18 20:54:03</td>\n",
       "      <td>161643.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>Success</td>\n",
       "      <td>512</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>cdeeaab9e1a7158375537018afb215d9</td>\n",
       "      <td>38fb9a0d17479f503be2d51270061280</td>\n",
       "      <td>54fdd12fff1505c8329c6fc50d4bcebc</td>\n",
       "      <td>c8a66d52ecfce75831966c472449e4dc</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>334b1cfe61c72a30619663a4dc44da49</td>\n",
       "      <td>compute_terminate_instance</td>\n",
       "      <td>2015-09-18 20:56:14</td>\n",
       "      <td>161774.0</td>\n",
       "      <td>2015-09-18 20:56:15</td>\n",
       "      <td>161775.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Success</td>\n",
       "      <td>512</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>cdeeaab9e1a7158375537018afb215d9</td>\n",
       "      <td>38fb9a0d17479f503be2d51270061280</td>\n",
       "      <td>54fdd12fff1505c8329c6fc50d4bcebc</td>\n",
       "      <td>c8a66d52ecfce75831966c472449e4dc</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>892a193b3818f7c9ff48e64c39e100a6</td>\n",
       "      <td>compute__do_build_and_run_instance</td>\n",
       "      <td>2015-09-18 20:56:33</td>\n",
       "      <td>161793.0</td>\n",
       "      <td>2015-09-18 20:56:38</td>\n",
       "      <td>161798.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>Success</td>\n",
       "      <td>512</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>cdeeaab9e1a7158375537018afb215d9</td>\n",
       "      <td>38fb9a0d17479f503be2d51270061280</td>\n",
       "      <td>54fdd12fff1505c8329c6fc50d4bcebc</td>\n",
       "      <td>c8a66d52ecfce75831966c472449e4dc</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>892a193b3818f7c9ff48e64c39e100a6</td>\n",
       "      <td>compute_terminate_instance</td>\n",
       "      <td>2015-09-18 21:01:55</td>\n",
       "      <td>162115.0</td>\n",
       "      <td>2015-09-18 21:01:56</td>\n",
       "      <td>162116.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Success</td>\n",
       "      <td>512</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>cdeeaab9e1a7158375537018afb215d9</td>\n",
       "      <td>38fb9a0d17479f503be2d51270061280</td>\n",
       "      <td>54fdd12fff1505c8329c6fc50d4bcebc</td>\n",
       "      <td>c8a66d52ecfce75831966c472449e4dc</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                      INSTANCE_UUID                               EVENT  \\\n",
       "0  467af885e36bbb3a656fb8f83f8c91db  compute__do_build_and_run_instance   \n",
       "1  84483962a31f776cf481fa56ff4babde  compute__do_build_and_run_instance   \n",
       "2  84483962a31f776cf481fa56ff4babde          compute_terminate_instance   \n",
       "3  974a749513018784b636dd1d6c2f435e  compute__do_build_and_run_instance   \n",
       "4  6ae938b5e39b74924f9a5a234d6fc8aa  compute__do_build_and_run_instance   \n",
       "5  6ae938b5e39b74924f9a5a234d6fc8aa          compute_terminate_instance   \n",
       "6  334b1cfe61c72a30619663a4dc44da49  compute__do_build_and_run_instance   \n",
       "7  334b1cfe61c72a30619663a4dc44da49          compute_terminate_instance   \n",
       "8  892a193b3818f7c9ff48e64c39e100a6  compute__do_build_and_run_instance   \n",
       "9  892a193b3818f7c9ff48e64c39e100a6          compute_terminate_instance   \n",
       "\n",
       "           START_TIME  START_SEC         FINISH_TIME  FINISH_SEC  \\\n",
       "0 2015-09-17 18:24:56    66296.0 2015-09-17 18:25:06     66306.0   \n",
       "1 2015-09-17 20:25:55    73555.0 2015-09-17 20:25:58     73558.0   \n",
       "2 2015-09-17 20:26:19    73579.0 2015-09-17 20:26:19     73579.0   \n",
       "3 2015-09-17 20:26:42    73602.0 2015-09-17 20:28:00     73680.0   \n",
       "4 2015-09-18 20:41:41   160901.0 2015-09-18 20:41:46    160906.0   \n",
       "5 2015-09-18 20:53:36   161616.0 2015-09-18 20:53:37    161617.0   \n",
       "6 2015-09-18 20:53:58   161638.0 2015-09-18 20:54:03    161643.0   \n",
       "7 2015-09-18 20:56:14   161774.0 2015-09-18 20:56:15    161775.0   \n",
       "8 2015-09-18 20:56:33   161793.0 2015-09-18 20:56:38    161798.0   \n",
       "9 2015-09-18 21:01:55   162115.0 2015-09-18 21:01:56    162116.0   \n",
       "\n",
       "   EVENT_DURATION   RESULT  MEMORY_MB  DISK_GB  VCPUS  \\\n",
       "0            10.0  Success        512        1      1   \n",
       "1             3.0  Success        512        1      1   \n",
       "2             0.0  Success        512        1      1   \n",
       "3            78.0  Success       4096       40      2   \n",
       "4             5.0  Success        512        1      1   \n",
       "5             1.0  Success        512        1      1   \n",
       "6             5.0  Success        512        1      1   \n",
       "7             1.0  Success        512        1      1   \n",
       "8             5.0  Success        512        1      1   \n",
       "9             1.0  Success        512        1      1   \n",
       "\n",
       "                      INSTANCE_NAME                           USER_ID  \\\n",
       "0  2c6824f29872b7ef0729973411589fc6  c358c86043c749283d1c498516e89b9a   \n",
       "1  fc5dac6c2f4d2a3223e4d298b3677740  c358c86043c749283d1c498516e89b9a   \n",
       "2  fc5dac6c2f4d2a3223e4d298b3677740  c358c86043c749283d1c498516e89b9a   \n",
       "3  fc5dac6c2f4d2a3223e4d298b3677740  c358c86043c749283d1c498516e89b9a   \n",
       "4  cdeeaab9e1a7158375537018afb215d9  38fb9a0d17479f503be2d51270061280   \n",
       "5  cdeeaab9e1a7158375537018afb215d9  38fb9a0d17479f503be2d51270061280   \n",
       "6  cdeeaab9e1a7158375537018afb215d9  38fb9a0d17479f503be2d51270061280   \n",
       "7  cdeeaab9e1a7158375537018afb215d9  38fb9a0d17479f503be2d51270061280   \n",
       "8  cdeeaab9e1a7158375537018afb215d9  38fb9a0d17479f503be2d51270061280   \n",
       "9  cdeeaab9e1a7158375537018afb215d9  38fb9a0d17479f503be2d51270061280   \n",
       "\n",
       "                         PROJECT_ID              HOST_NAME (PHYSICAL)  \n",
       "0  54fdd12fff1505c8329c6fc50d4bcebc  40819f1d02e0ce5401227d282b5f58ad  \n",
       "1  54fdd12fff1505c8329c6fc50d4bcebc  2412e43489814bc20cb5382e66e4d1f0  \n",
       "2  54fdd12fff1505c8329c6fc50d4bcebc  2412e43489814bc20cb5382e66e4d1f0  \n",
       "3  54fdd12fff1505c8329c6fc50d4bcebc  2412e43489814bc20cb5382e66e4d1f0  \n",
       "4  54fdd12fff1505c8329c6fc50d4bcebc  c8a66d52ecfce75831966c472449e4dc  \n",
       "5  54fdd12fff1505c8329c6fc50d4bcebc  c8a66d52ecfce75831966c472449e4dc  \n",
       "6  54fdd12fff1505c8329c6fc50d4bcebc  c8a66d52ecfce75831966c472449e4dc  \n",
       "7  54fdd12fff1505c8329c6fc50d4bcebc  c8a66d52ecfce75831966c472449e4dc  \n",
       "8  54fdd12fff1505c8329c6fc50d4bcebc  c8a66d52ecfce75831966c472449e4dc  \n",
       "9  54fdd12fff1505c8329c6fc50d4bcebc  c8a66d52ecfce75831966c472449e4dc  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# show instance events\n",
    "instance_events_df.head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Favorite Flavors\n",
    "\n",
    "In OpenStack, flavors define the compute, memory, and storage capacity of nova computing instances. To put it simply, a flavor is an available hardware configuration for a server. It defines the size of a virtual server that can be launched. [\\[1\\]][1].\n",
    "\n",
    "[1]: https://docs.openstack.org/horizon/latest/admin/manage-flavors.html\n",
    "\n",
    "### Instance flavors available at Chameleon KVM@TACC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>FLAVOR</th>\n",
       "      <th>VCPUS</th>\n",
       "      <th>MEMORY_MB</th>\n",
       "      <th>DISK_GB</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>m1.tiny</td>\n",
       "      <td>1</td>\n",
       "      <td>512</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>m1.small</td>\n",
       "      <td>1</td>\n",
       "      <td>2048</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>m1.medium</td>\n",
       "      <td>2</td>\n",
       "      <td>4096</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>storage.medium</td>\n",
       "      <td>1</td>\n",
       "      <td>4096</td>\n",
       "      <td>2048</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>m1.large</td>\n",
       "      <td>4</td>\n",
       "      <td>8192</td>\n",
       "      <td>80</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>m1.xlarge</td>\n",
       "      <td>8</td>\n",
       "      <td>16384</td>\n",
       "      <td>160</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>m1.xxlarge</td>\n",
       "      <td>8</td>\n",
       "      <td>32768</td>\n",
       "      <td>160</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>m1.xxxlarge</td>\n",
       "      <td>16</td>\n",
       "      <td>32768</td>\n",
       "      <td>160</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           FLAVOR  VCPUS  MEMORY_MB  DISK_GB\n",
       "0         m1.tiny      1        512        1\n",
       "1        m1.small      1       2048       20\n",
       "2       m1.medium      2       4096       40\n",
       "3  storage.medium      1       4096     2048\n",
       "4        m1.large      4       8192       80\n",
       "5       m1.xlarge      8      16384      160\n",
       "6      m1.xxlarge      8      32768      160\n",
       "7     m1.xxxlarge     16      32768      160"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "flavors = [\n",
    "    {\n",
    "        'FLAVOR': 'm1.tiny',\n",
    "        'VCPUS': 1,\n",
    "        'DISK_GB': 1,\n",
    "        'MEMORY_MB': 512\n",
    "    },\n",
    "    {\n",
    "        'FLAVOR': 'm1.small',\n",
    "        'VCPUS': 1,\n",
    "        'DISK_GB': 20,\n",
    "        'MEMORY_MB': 2048\n",
    "    },\n",
    "    {\n",
    "        'FLAVOR': 'm1.medium',\n",
    "        'VCPUS': 2,\n",
    "        'DISK_GB': 40,\n",
    "        'MEMORY_MB': 4096\n",
    "    },\n",
    "    {\n",
    "        'FLAVOR': 'storage.medium',\n",
    "        'VCPUS': 1,\n",
    "        'DISK_GB': 2048,\n",
    "        'MEMORY_MB': 4096\n",
    "    },\n",
    "    {\n",
    "        'FLAVOR': 'm1.large',\n",
    "        'VCPUS': 4,\n",
    "        'DISK_GB': 80,\n",
    "        'MEMORY_MB': 8192\n",
    "    },\n",
    "    {\n",
    "        'FLAVOR': 'm1.xlarge',\n",
    "        'VCPUS': 8,\n",
    "        'DISK_GB': 160,\n",
    "        'MEMORY_MB': 16384\n",
    "    },\n",
    "    {\n",
    "        'FLAVOR': 'm1.xxlarge',\n",
    "        'VCPUS': 8,\n",
    "        'DISK_GB': 160,\n",
    "        'MEMORY_MB': 32768\n",
    "    },\n",
    "    {\n",
    "        'FLAVOR': 'm1.xxxlarge',\n",
    "        'VCPUS': 16,\n",
    "        'DISK_GB': 160,\n",
    "        'MEMORY_MB': 32768\n",
    "    }\n",
    "]\n",
    "flavors_df = pd.DataFrame(flavors)[['FLAVOR', 'VCPUS', 'MEMORY_MB', 'DISK_GB']]\n",
    "flavors_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# function for plotting KVM type\n",
    "def plot_bar_percent(df, figsize, fontsize):\n",
    "    ax = df.plot(kind='bar', figsize=figsize)\n",
    "    \n",
    "    # create a list to collect the plt.patches data\n",
    "    totals = []\n",
    "    # find the values and append to list\n",
    "    for i in ax.patches:\n",
    "        totals.append(i.get_height())\n",
    "\n",
    "    # set individual bar lables using above list\n",
    "    total = sum(totals)\n",
    "    for i in ax.patches:\n",
    "        # get_x pulls left or right; get_height pushes up or down\n",
    "        ax.text(i.get_x()-.01, i.get_height()*1.02, str(i.get_height()) + ' (' + str(round((i.get_height()/total)*100, 2))+'%)', fontsize=fontsize)\n",
    "    \n",
    "\n",
    "kvm_cloud_traces_flavor_df = pd.merge(instance_events_df, flavors_df,  how='left', left_on=['VCPUS','MEMORY_MB','DISK_GB'], right_on = ['VCPUS','MEMORY_MB','DISK_GB'])\n",
    "\n",
    "instance_type_df = kvm_cloud_traces_flavor_df.groupby('INSTANCE_UUID').first().reset_index()\n",
    "\n",
    "plot_bar_percent(instance_type_df[['INSTANCE_UUID', 'FLAVOR']].groupby('FLAVOR').size().to_frame('COUNT').sort_values(by=['COUNT'], ascending=False), (20,7), 15)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}