example.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Yardstick Example\n",
    "\n",
    "This example notebook provides a simple example of how you can use the Yardstick benchmark to collect performance metrics from Minecraft-like games and how to evaluate their performance."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Running an experiment\n",
    "\n",
    "The cell below shows you how to run a simple experiment."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "PLAY [Clean data from nodes] ***************************************************\n",
      "\n",
      "TASK [Gathering Facts] *********************************************************\n",
      "\u001b[0;32mok: [node015]\u001b[0m\n",
      "\u001b[0;32mok: [node001]\u001b[0m\n",
      "\n",
      "TASK [Remove data from nodes] **************************************************\n",
      "\u001b[0;32mok: [node001]\u001b[0m\n",
      "\u001b[0;32mok: [node015]\u001b[0m\n",
      "\n",
      "PLAY RECAP *********************************************************************\n",
      "\u001b[0;32mnode001\u001b[0m                    : \u001b[0;32mok=2   \u001b[0m changed=0    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   \n",
      "\u001b[0;32mnode015\u001b[0m                    : \u001b[0;32mok=2   \u001b[0m changed=0    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   \n",
      "\n",
      "PLAY [Deploy Telegraf] *********************************************************\n",
      "\n",
      "TASK [Gathering Facts] *********************************************************\n",
      "\u001b[0;32mok: [node015]\u001b[0m\n",
      "\u001b[0;32mok: [node001]\u001b[0m\n",
      "\n",
      "TASK [Create Telegraf directory] ***********************************************\n",
      "\u001b[0;33mchanged: [node001]\u001b[0m\n",
      "\u001b[0;33mchanged: [node015]\u001b[0m\n",
      "\n",
      "TASK [Download Telegraf] *******************************************************\n",
      "\u001b[0;33mchanged: [node001]\u001b[0m\n",
      "\u001b[0;33mchanged: [node015]\u001b[0m\n",
      "\n",
      "TASK [Copy Telegraf config] ****************************************************\n",
      "\u001b[0;33mchanged: [node001]\u001b[0m\n",
      "\u001b[0;33mchanged: [node015]\u001b[0m\n",
      "\n",
      "PLAY [Deploy Minecraft tick collection execd input script] *********************\n",
      "\n",
      "TASK [Gathering Facts] *********************************************************\n",
      "\u001b[0;32mok: [node001]\u001b[0m\n",
      "\n",
      "TASK [Copy Minecraft tick collection input script] *****************************\n",
      "\u001b[0;33mchanged: [node001]\u001b[0m\n",
      "\n",
      "PLAY RECAP *********************************************************************\n",
      "\u001b[0;33mnode001\u001b[0m                    : \u001b[0;32mok=6   \u001b[0m \u001b[0;33mchanged=4   \u001b[0m unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   \n",
      "\u001b[0;33mnode015\u001b[0m                    : \u001b[0;32mok=4   \u001b[0m \u001b[0;33mchanged=3   \u001b[0m unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   \n",
      "\n",
      "PLAY [Start Telegraf] **********************************************************\n",
      "\n",
      "TASK [Gathering Facts] *********************************************************\n",
      "\u001b[0;32mok: [node015]\u001b[0m\n",
      "\u001b[0;32mok: [node001]\u001b[0m\n",
      "\n",
      "TASK [Find executable] *********************************************************\n",
      "\u001b[0;33mchanged: [node001]\u001b[0m\n",
      "\u001b[0;33mchanged: [node015]\u001b[0m\n",
      "\n",
      "TASK [Run Telegraf] ************************************************************\n",
      "\u001b[0;33mchanged: [node001]\u001b[0m\n",
      "\u001b[0;33mchanged: [node015]\u001b[0m\n",
      "\n",
      "PLAY RECAP *********************************************************************\n",
      "\u001b[0;33mnode001\u001b[0m                    : \u001b[0;32mok=3   \u001b[0m \u001b[0;33mchanged=2   \u001b[0m unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   \n",
      "\u001b[0;33mnode015\u001b[0m                    : \u001b[0;32mok=3   \u001b[0m \u001b[0;33mchanged=2   \u001b[0m unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   \n",
      "\n",
      "PLAY [Deploy PaperMC] **********************************************************\n",
      "\n",
      "TASK [Gathering Facts] *********************************************************\n",
      "\u001b[0;32mok: [node001]\u001b[0m\n",
      "\n",
      "TASK [file] ********************************************************************\n",
      "\u001b[0;33mchanged: [node001]\u001b[0m\n",
      "\n",
      "TASK [Download PaperMC] ********************************************************\n",
      "\u001b[0;33mchanged: [node001]\u001b[0m\n",
      "\n",
      "TASK [Copy config file] ********************************************************\n",
      "\u001b[0;33mchanged: [node001]\u001b[0m\n",
      "\n",
      "TASK [Accept EULA] *************************************************************\n",
      "\u001b[0;33mchanged: [node001]\u001b[0m\n",
      "\n",
      "TASK [Download Jolokia JVM Agent] **********************************************\n",
      "\u001b[0;33mchanged: [node001]\u001b[0m\n",
      "\n",
      "PLAY RECAP *********************************************************************\n",
      "\u001b[0;33mnode001\u001b[0m                    : \u001b[0;32mok=6   \u001b[0m \u001b[0;33mchanged=5   \u001b[0m unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   \n",
      "\n",
      "PLAY [Deploy PaperMC] **********************************************************\n",
      "\n",
      "TASK [Gathering Facts] *********************************************************\n",
      "\u001b[0;32mok: [node001]\u001b[0m\n",
      "\n",
      "TASK [Run PaperMC] *************************************************************\n",
      "\u001b[0;33mchanged: [node001]\u001b[0m\n",
      "\n",
      "TASK [Waiting for server to become ready] **************************************\n",
      "\u001b[0;32mok: [node001]\u001b[0m\n",
      "\n",
      "PLAY RECAP *********************************************************************\n",
      "\u001b[0;33mnode001\u001b[0m                    : \u001b[0;32mok=3   \u001b[0m \u001b[0;33mchanged=1   \u001b[0m unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   \n",
      "\n",
      "PLAY [Deploy Minecraft WalkAround workload] ************************************\n",
      "\n",
      "TASK [Gathering Facts] *********************************************************\n",
      "\u001b[0;32mok: [node015]\u001b[0m\n",
      "\n",
      "TASK [debug] *******************************************************************\n",
      "\u001b[0;32mok: [node015] => {\u001b[0m\n",
      "\u001b[0;32m    \"msg\": \"working directory is: /local/core2435/yardstick/node015/walkaround-ta7vf2a2\"\u001b[0m\n",
      "\u001b[0;32m}\u001b[0m\n",
      "\n",
      "TASK [file] ********************************************************************\n",
      "\u001b[0;33mchanged: [node015]\u001b[0m\n",
      "\n",
      "TASK [Check if nvm is available] ***********************************************\n",
      "\u001b[0;33mchanged: [node015]\u001b[0m\n",
      "\n",
      "TASK [Install nvm] *************************************************************\n",
      "\u001b[0;33mchanged: [node015]\u001b[0m\n",
      "\n",
      "TASK [Install node 18 and Mineflayer] ******************************************\n",
      "\u001b[0;33mchanged: [node015]\u001b[0m\n",
      "\n",
      "TASK [Checkout RCON client] ****************************************************\n",
      "\u001b[0;33mchanged: [node015]\u001b[0m\n",
      "\n",
      "TASK [Copy scripts] ************************************************************\n",
      "\u001b[0;33mchanged: [node015] => (item=/var/scratch/core2435/miniconda3/envs/yardstick/lib/python3.9/site-packages/yardstick_benchmark/games/minecraft/workload/set_spawn.js)\u001b[0m\n",
      "\u001b[0;33mchanged: [node015] => (item=/var/scratch/core2435/miniconda3/envs/yardstick/lib/python3.9/site-packages/yardstick_benchmark/games/minecraft/workload/walkaround_bot.js)\u001b[0m\n",
      "\u001b[0;33mchanged: [node015] => (item=/var/scratch/core2435/miniconda3/envs/yardstick/lib/python3.9/site-packages/yardstick_benchmark/games/minecraft/workload/walkaround_worker_bot.js)\u001b[0m\n",
      "\n",
      "PLAY RECAP *********************************************************************\n",
      "\u001b[0;33mnode015\u001b[0m                    : \u001b[0;32mok=8   \u001b[0m \u001b[0;33mchanged=6   \u001b[0m unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   \n",
      "\n",
      "PLAY [all] *********************************************************************\n",
      "\n",
      "TASK [Gathering Facts] *********************************************************\n",
      "\u001b[0;32mok: [node015]\u001b[0m\n",
      "\n",
      "TASK [Set game spawn location] *************************************************\n",
      "\u001b[0;33mchanged: [node015]\u001b[0m\n",
      "\n",
      "TASK [Run Minecraft bot] *******************************************************\n",
      "\u001b[0;33mchanged: [node015]\u001b[0m\n",
      "\n",
      "PLAY RECAP *********************************************************************\n",
      "\u001b[0;33mnode015\u001b[0m                    : \u001b[0;32mok=3   \u001b[0m \u001b[0;33mchanged=2   \u001b[0m unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   \n",
      "sleeping for 10 seconds\n",
      "\n",
      "PLAY [Stop PaperMC] ************************************************************\n",
      "\n",
      "TASK [Gathering Facts] *********************************************************\n",
      "\u001b[0;32mok: [node001]\u001b[0m\n",
      "\n",
      "TASK [Find PaperMC PID] ********************************************************\n",
      "\u001b[0;33mchanged: [node001]\u001b[0m\n",
      "\n",
      "TASK [Obtain PaperMC pid] ******************************************************\n",
      "\u001b[0;32mok: [node001]\u001b[0m\n",
      "\n",
      "TASK [Stop PaperMC] ************************************************************\n",
      "\u001b[0;33mchanged: [node001]\u001b[0m\n",
      "\n",
      "TASK [wait_for] ****************************************************************\n",
      "\u001b[0;32mok: [node001]\u001b[0m\n",
      "\n",
      "PLAY RECAP *********************************************************************\n",
      "\u001b[0;33mnode001\u001b[0m                    : \u001b[0;32mok=5   \u001b[0m \u001b[0;33mchanged=2   \u001b[0m unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   \n",
      "\u001b[0;31mERROR! Empty playbook, nothing to do: /var/scratch/core2435/miniconda3/envs/yardstick/lib/python3.9/site-packages/yardstick_benchmark/games/minecraft/server/papermc_cleanup.yml\u001b[0m\n",
      "\n",
      "PLAY [Stop Telegraf] ***********************************************************\n",
      "\n",
      "TASK [Gathering Facts] *********************************************************\n",
      "\u001b[0;32mok: [node015]\u001b[0m\n",
      "\u001b[0;32mok: [node001]\u001b[0m\n",
      "\n",
      "TASK [Lookup Telegraf PID] *****************************************************\n",
      "\u001b[0;33mchanged: [node001]\u001b[0m\n",
      "\u001b[0;33mchanged: [node015]\u001b[0m\n",
      "\n",
      "TASK [set_fact] ****************************************************************\n",
      "\u001b[0;32mok: [node001]\u001b[0m\n",
      "\u001b[0;32mok: [node015]\u001b[0m\n",
      "\n",
      "TASK [Stop Telegraf] ***********************************************************\n",
      "\u001b[0;33mchanged: [node015]\u001b[0m\n",
      "\u001b[0;33mchanged: [node001]\u001b[0m\n",
      "\n",
      "TASK [wait_for] ****************************************************************\n",
      "\u001b[0;32mok: [node001]\u001b[0m\n",
      "\u001b[0;32mok: [node015]\u001b[0m\n",
      "\n",
      "PLAY RECAP *********************************************************************\n",
      "\u001b[0;33mnode001\u001b[0m                    : \u001b[0;32mok=5   \u001b[0m \u001b[0;33mchanged=2   \u001b[0m unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   \n",
      "\u001b[0;33mnode015\u001b[0m                    : \u001b[0;32mok=5   \u001b[0m \u001b[0;33mchanged=2   \u001b[0m unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   \n",
      "\u001b[0;31mERROR! Empty playbook, nothing to do: /var/scratch/core2435/miniconda3/envs/yardstick/lib/python3.9/site-packages/yardstick_benchmark/monitoring/telegraf_cleanup.yml\u001b[0m\n",
      "\n",
      "PLAY [Get data from nodes] *****************************************************\n",
      "\n",
      "TASK [Gathering Facts] *********************************************************\n",
      "\u001b[0;32mok: [node015]\u001b[0m\n",
      "\u001b[0;32mok: [node001]\u001b[0m\n",
      "\n",
      "TASK [Get data from nodes] *****************************************************\n",
      "\u001b[0;35m[DEPRECATION WARNING]: The connection's stdin object is deprecated. Call \u001b[0m\n",
      "\u001b[0;35mdisplay.prompt_until(msg) instead. This feature will be removed in version \u001b[0m\n",
      "\u001b[0;35m2.19. Deprecation warnings can be disabled by setting \u001b[0m\n",
      "\u001b[0;35mdeprecation_warnings=False in ansible.cfg.\u001b[0m\n",
      "\u001b[0;35m[DEPRECATION WARNING]: The connection's stdin object is deprecated. Call \u001b[0m\n",
      "\u001b[0;35mdisplay.prompt_until(msg) instead. This feature will be removed in version \u001b[0m\n",
      "\u001b[0;35m2.19. Deprecation warnings can be disabled by setting \u001b[0m\n",
      "\u001b[0;35mdeprecation_warnings=False in ansible.cfg.\u001b[0m\n",
      "\u001b[0;33mchanged: [node015]\u001b[0m\n",
      "\u001b[0;33mchanged: [node001]\u001b[0m\n",
      "\n",
      "TASK [debug] *******************************************************************\n",
      "\u001b[0;32mok: [node001] => {\u001b[0m\n",
      "\u001b[0;32m    \"msg\": \"Output files have been written to /var/scratch/core2435/yardstick/output\"\u001b[0m\n",
      "\u001b[0;32m}\u001b[0m\n",
      "\u001b[0;32mok: [node015] => {\u001b[0m\n",
      "\u001b[0;32m    \"msg\": \"Output files have been written to /var/scratch/core2435/yardstick/output\"\u001b[0m\n",
      "\u001b[0;32m}\u001b[0m\n",
      "\n",
      "PLAY RECAP *********************************************************************\n",
      "\u001b[0;33mnode001\u001b[0m                    : \u001b[0;32mok=3   \u001b[0m \u001b[0;33mchanged=1   \u001b[0m unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   \n",
      "\u001b[0;33mnode015\u001b[0m                    : \u001b[0;32mok=3   \u001b[0m \u001b[0;33mchanged=1   \u001b[0m unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   \n",
      "\n",
      "PLAY [Clean data from nodes] ***************************************************\n",
      "\n",
      "TASK [Gathering Facts] *********************************************************\n",
      "\u001b[0;32mok: [node015]\u001b[0m\n",
      "\u001b[0;32mok: [node001]\u001b[0m\n",
      "\n",
      "TASK [Remove data from nodes] **************************************************\n",
      "\u001b[0;33mchanged: [node001]\u001b[0m\n",
      "\u001b[0;33mchanged: [node015]\u001b[0m\n",
      "\n",
      "PLAY RECAP *********************************************************************\n",
      "\u001b[0;33mnode001\u001b[0m                    : \u001b[0;32mok=2   \u001b[0m \u001b[0;33mchanged=1   \u001b[0m unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   \n",
      "\u001b[0;33mnode015\u001b[0m                    : \u001b[0;32mok=2   \u001b[0m \u001b[0;33mchanged=1   \u001b[0m unreachable=0    failed=0    skipped=0    rescued=0    ignored=0   \n"
     ]
    }
   ],
   "source": [
    "from yardstick_benchmark.provisioning import Das\n",
    "from yardstick_benchmark.monitoring import Telegraf\n",
    "from yardstick_benchmark.games.minecraft.server import PaperMC\n",
    "from yardstick_benchmark.games.minecraft.workload import WalkAround\n",
    "import yardstick_benchmark\n",
    "from time import sleep\n",
    "from datetime import datetime\n",
    "from pathlib import Path\n",
    "import os\n",
    "import shutil\n",
    "\n",
    "dest = Path(f\"/var/scratch/{os.getlogin()}/yardstick/output\")\n",
    "if dest.exists():\n",
    "    shutil.rmtree(dest)\n",
    "\n",
    "### DEPLOYMENT ENVIRONMENT ###\n",
    "\n",
    "# The DAS compute cluster is a medium-sized cluster for research and education.\n",
    "# We use it in this example to provision bare-metal machines to run our performance\n",
    "# evaluation.\n",
    "das = Das()\n",
    "# We reserve 2 nodes.\n",
    "nodes = das.provision(num=2)\n",
    "\n",
    "try:\n",
    "    # Just in case, we remove data that may have been left from a previous run.\n",
    "    yardstick_benchmark.clean(nodes)\n",
    "\n",
    "    ### METRICS ###\n",
    "\n",
    "    # Telegraf (https://www.influxdata.com/time-series-platform/telegraf/)\n",
    "    # is the metric collection tool we use to collect performance metrics from the\n",
    "    # nodes and any applications deployed on these nodes.\n",
    "    telegraf = Telegraf(nodes)\n",
    "    # We plan to deploy our Minecraft-like game server on node 0.\n",
    "    # To obtain application level metrics from the game server,\n",
    "    # the next two lines configure node 0 to run additional metric collection\n",
    "    # tools.\n",
    "    telegraf.add_input_jolokia_agent(nodes[0])\n",
    "    telegraf.add_input_execd_minecraft_ticks(nodes[0])\n",
    "    # Perform the actual deployment of Telegraf.\n",
    "    # This includes downloading the Telegraf executable and preparing configuration\n",
    "    # files.\n",
    "    res = telegraf.deploy()\n",
    "    # Start Telegraf on all remote nodes.\n",
    "    telegraf.start()\n",
    "\n",
    "    ### System Under Test (SUT) ###\n",
    "\n",
    "    # PaperMC (https://papermc.io/) is the Minecraft-like game whose performance\n",
    "    # we'll evaluate in this example.\n",
    "    # We pass a list with all the nodes on which we want to deploy a server.\n",
    "    # In this example, we only deploy a server on node 0.\n",
    "    papermc = PaperMC(nodes[:1])\n",
    "    # We perform the deployment, including downloading the game executable JAR and\n",
    "    # correctly configuring the game's configuration file.\n",
    "    papermc.deploy()\n",
    "    # We start the game server.\n",
    "    papermc.start()\n",
    "\n",
    "    ### WORKLOAD ###\n",
    "\n",
    "    wl = WalkAround(nodes[1:], nodes[0].host, bots_per_node=10)\n",
    "    wl.deploy()\n",
    "    wl.start()\n",
    "\n",
    "    sleep_time = 60\n",
    "    print(f\"sleeping for {sleep_time} seconds\")\n",
    "    sleep(sleep_time)\n",
    "\n",
    "    papermc.stop()\n",
    "    papermc.cleanup()\n",
    "\n",
    "    telegraf.stop()\n",
    "    telegraf.cleanup()\n",
    "\n",
    "    yardstick_benchmark.fetch(dest, nodes)\n",
    "finally:\n",
    "    yardstick_benchmark.clean(nodes)\n",
    "    das.release(nodes)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Pre-Processing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'dest' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[1], line 4\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m \u001b[38;5;21;01mglob\u001b[39;00m\n\u001b[1;32m      2\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m \u001b[38;5;21;01mpandas\u001b[39;00m \u001b[38;5;28;01mas\u001b[39;00m \u001b[38;5;21;01mpd\u001b[39;00m\n\u001b[0;32m----> 4\u001b[0m raw_data_files \u001b[38;5;241m=\u001b[39m glob\u001b[38;5;241m.\u001b[39mglob(\u001b[38;5;124mf\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;132;01m{\u001b[39;00mdest\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m/**/metrics-*.csv\u001b[39m\u001b[38;5;124m\"\u001b[39m, recursive\u001b[38;5;241m=\u001b[39m\u001b[38;5;28;01mTrue\u001b[39;00m)\n\u001b[1;32m      5\u001b[0m \u001b[38;5;28mprint\u001b[39m(raw_data_files)\n\u001b[1;32m      7\u001b[0m \u001b[38;5;28;01mfor\u001b[39;00m raw_data_file \u001b[38;5;129;01min\u001b[39;00m raw_data_files:\n",
      "\u001b[0;31mNameError\u001b[0m: name 'dest' is not defined"
     ]
    }
   ],
   "source": [
    "import glob\n",
    "import pandas as pd\n",
    "\n",
    "raw_data_files = glob.glob(f\"{dest}/**/metrics-*.csv\", recursive=True)\n",
    "print(raw_data_files)\n",
    "\n",
    "for raw_data_file in raw_data_files:\n",
    "    metrics_file = Path(raw_data_file)\n",
    "    keys = {}\n",
    "    with open(metrics_file) as fin:\n",
    "        for line in fin:\n",
    "            first_delim = line.find(\",\")\n",
    "            second_delim = line.find(\",\", first_delim+1)\n",
    "            key = line[first_delim+1:second_delim]\n",
    "            if key not in keys:\n",
    "                keys[key] = open(metrics_file.parent / f\"{key}.csv\", \"w+\")\n",
    "            keys[key].write(line)\n",
    "    for key, fd in keys.items():\n",
    "        fd.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Visualizing Results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: xlabel='Time [m]', ylabel='CPU utilization [%]'>"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import seaborn as sns\n",
    "\n",
    "dfs = []\n",
    "for cpu_file in glob.glob(f\"{dest}/**/cpu.csv\", recursive=True):\n",
    "    df = pd.read_csv(cpu_file, names = [\"timestamp\",\"measurement\",\"core_id\",\"cpu\",\"host\",\"physical_id\",\"time_active\",\"time_guest\",\"time_guest_nice\",\"time_idle\",\"time_iowait\",\"time_irq\",\"time_nice\",\"time_softirq\",\"time_steal\",\"time_system\",\"time_user\"])\n",
    "    df[\"node\"] = Path(cpu_file).parent.parent.name\n",
    "    df[\"timestamp\"] = df[\"timestamp\"].transform(lambda x: x - x.min())\n",
    "    df[\"timestamp_m\"] = df[\"timestamp\"] / 60\n",
    "    df = df[df.cpu == \"cpu-total\"]\n",
    "    df['time_total'] = df.time_active + df.time_idle\n",
    "    df['util'] = 100 * df.time_active / df.time_total\n",
    "    df = df[df.groupby([\"timestamp\", \"cpu\"])[\"time_total\"].rank(ascending=False) <= 1]\n",
    "    dfs.append(df)\n",
    "df = pd.concat(dfs, ignore_index=True)\n",
    "df.head()\n",
    "\n",
    "custom_params = {\"axes.spines.right\": False, \"axes.spines.top\": False}\n",
    "sns.set_theme(style=\"ticks\", rc=custom_params)\n",
    "ax = sns.lineplot(df, x=\"timestamp_m\", y=\"util\", hue=\"node\")\n",
    "ax.grid(axis=\"y\")\n",
    "ax.set_ylim(bottom=0)\n",
    "ax.set_ylabel(\"CPU utilization [%]\")\n",
    "ax.set_xlabel(\"Time [m]\")\n",
    "ax\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: xlabel='Time [m]', ylabel='Tick duration [ms]'>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# timestamp,measurement,cpu,host,usage_guest,usage_guest_nice,usage_idle,usage_iowait,usage_irq,usage_nice,usage_softirq,usage_steal,usage_system,usage_user\n",
    "tick_times_file = glob.glob(f\"{dest}/**/minecraft_tick_times.csv\", recursive=True)[0]\n",
    "\n",
    "df = pd.read_csv(tick_times_file, names = [\"timestamp\", \"label\", \"node\", \"jolokia_endpoint\", \"tick_duration_ms\"])\n",
    "df[\"timestamp\"] = df[\"timestamp\"].transform(lambda x: x - x.min())\n",
    "df[\"timestamp_m\"] = df[\"timestamp\"] / 60\n",
    "df.head()\n",
    "\n",
    "custom_params = {\"axes.spines.right\": False, \"axes.spines.top\": False}\n",
    "sns.set_theme(style=\"ticks\", rc=custom_params)\n",
    "ax = sns.lineplot(df, x=\"timestamp_m\", y=\"tick_duration_ms\")\n",
    "ax.grid(axis=\"y\")\n",
    "ax.set_ylim(bottom=0)\n",
    "ax.set_ylabel(\"Tick duration [ms]\")\n",
    "ax.set_xlabel(\"Time [m]\")\n",
    "ax"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "pamuk-QFfrVgWj-py3.9",
   "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": "3.9.19"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}