@pathable/meteor-galaxy-auto-scaling

Meteor Galaxy Auto Scaling (mgas)

NodeJS command line tool to monitor and auto-scale Meteor Galaxy

⚠️ Important: we are already using it in production for at least two months but please monitor your containers to be sure everything is behaving as expected and please report any issue. ⚠️

Features

• Monitoring: reads Galaxy and Meteor APM metrics
• Alerts: reports containers and apps not running as expected
• Auto-scaling: take actions when specific conditions are met
• Send slack notifications: updates, alerts and autos-scaling messages;

How it works

We use puppeteer to read Galaxy and Meteor APM dashboards and then we execute actions based on your desired settings.

How to use

Set up in your CI server or equivalent a job to call mgas (meteor-galaxy-auto-scaling) command from time to time, it's probably a good idea to use a very short period, like every 2 minutes then you will be always monitoring your containers.

If you need help configuring your CI check our configurations examples. Please open an issue if you are having a hard time then we can improve the documentation. If you have already configured your CI please open a Pull Request including the instructions in the configuration examples.

First you need to install mgas (meteor-galaxy-auto-scaling) using yarn or npm:

yarn global add @pathable/meteor-galaxy-auto-scaling

npm install -g @pathable/meteor-galaxy-auto-scaling

Then you run it informing a configuration file with your alerts and auto-scaling rules:

mgas --settings settings.json

You can have different settings for different purposes.

Updates

Check the log of changes.

Settings

{
  "appName": "your app host (required)",
  "username": "your Galaxy username (required)",
  "password": "your Galaxy password (required)",
  "slackWebhook": "your Slack webhook URL",
  "silentSlack": false,
  "simulation": false,  
  "persistentStorage": "full path to where we want to storage scrapped info",
  "infoRules": {
    "send": true,   
    "channel": "#galaxy-updates"
  },
  "alertRules": {
    "channel": "#alerts",
    "messagePrefix": "@channel",
    "maxInContainers": {
    },
    "maxInApp": {
      "pubSubResponseTime": 200,
      "methodResponseTime": 300,
      "cpuUsageAverage": 1,
      "memoryUsageByHost": 1500,
      "sessionsByHost": 5
    }
  },
  "autoscaleRules": {
    "channel": "#auto-scaling",
    "containersToScale": 2,
    "minContainers": 2,
    "maxContainers": 10,
    "addWhen": {
      "pubSubResponseTimeAbove": 300,
      "methodResponseTimeAbove": 300,
      "cpuAbove": 50,
      "memoryAbove":  70,
      "sessionsAbove":  50
    },
    "reduceWhen": {
      "pubSubResponseTimeBelow": 300,
      "methodResponseTimeBelow": 300,
      "cpuBelow": 25,
      "memoryBelow": 25,
      "sessionsBelow": 30
    }
  },
  "minimumStats": 5,
  "puppeteer": {
    "headless": true
  }
}

Auto scale rules

The autoscaling (autoscaleRules) behavior is meant to adjust smartly the containers on the Galaxy server taking into account the data got from there and a predefined configuration.

• Three actions are supported:

  • add containers (conditions are configured on addWhen json key);
  • reduce containers (conditions are configured on reduceWhen json key);

• The conditions available are: "[pubSubResponseTime|methodResponseTime|cpu|memory|sessions][Above|Below]". Check out to which values refer for each: from Galaxy Panel and APM panel.

• The conditions express the property average on the active containers. The active containers are those that are running, the ones starting or stopping are ignored.

• Multiple conditions can be informed and they are evaluated in different ways depending on the action

  • add action evaluates with OR, one match is enough to add new container
  • reduce action evaluates with AND, one miss match is enough to not remove one container

• The addWhen and reduceWhen behaviors check to not go beyond a containers count range. This range is described by the minContainers and maxContainers configuration.

• The addWhen and reduceWhen behaviors won't run if a scaling is happening. If any other condition passes it will run on the next run.

• An slack message is sent anytime a scaling behavior is triggered if you set a Slack Webhook, the messages are sent to the default webhook channel. You will receive messages like this

Alerts

You can set maximum limits for container metrics (CPU, memory and connected clients) and also for Meteor app metrics (response time for publications and methods).

Example:

"alertRules": {
    "maxInContainers": {
      "cpu": 1,
      "memory": 10,
      "clients": 5
    },
    "maxInApp": {
      "pubSubResponseTime": 200,
      "methodResponseTime": 300
    }
},
"minimumStats": 5,

You will receive an alert like this when at least minimumStats times in a row the current value was above the maximum expected.

Then if you run mgas every 2 minutes and use minimumStatus as 5 you will get an alert when your metric is at least 10 minutes above the maximum expected .

minimumStats is set in the first level of the settings because maybe we will use this information in the future also for auto-scaling, for now auto-scaling is not considering the minimumStats value.

Info rules

• Set the channel, by default will go to default webhook channel
• You will receive messages like this

Advanced

Remote settings

If you need to have dynamic settings coming from a external location, like an API, you can configure:

"remote": {
   "url": "https://yourapi.com/v1/auto-scaling?yourkey=XXX&anySetting=YYY"
}

Then the JSON returned by this API will be merged (using lodash.merge) with your local settings . If the request to this URL throws an error then the local settings will be used anyway and a console.error (Error getting remote options from ${url}) will be printed.

Developing

If you want to include new features that includes reading new data from Galaxy or Meteor APM you will probably want to run puppeteer watching the actions, then change headless setting to false.

  "puppeteer": {
    "headless": false
  }

Troubleshooting

Fixing Puppeteer on Ubuntu 16.04

sudo apt-get install libx11-xcb1 libxcomposite1 libxi6 libxext6 libxtst6 libnss3 libcups2 libxss1 libxrandr2 libasound2 libpangocairo-1.0-0 libatk1.0-0 libatk-bridge2.0-0 libgtk-3-0

Contributions

Please open issues to discuss improvements and report bugs. Also feel free to submit PRs, it is always a good idea to discuss first your PR idea in the issues.

Contributors ✨

Thanks goes to these wonderful people (emoji key):

Filipe Névola 💻

Nacho Codoñer 💻

This project follows the all-contributors specification. Contributions of any kind welcome!

LICENSE

MIT