Guilty Spark

"Greetings. I am the Monitor of Installation 04. I am 343 Guilty Spark."
- 343 Guilty Spark

Easy, cheap monitoring for your personal sites. Minimal set up. Uses twilio for messaging.

I am running it through cron every hour.

Set Up

Websites and Contacts

There are two files that aren't provided for obvious reasons:

• websites.json
• contacts.json

websites.json should be a list of json objects that represent the website to monitor. A website object has the following properties:

field	description	required
url	url of the website	yes
displayName	a display name to reference the website for logs	yes

Example websites.json file:

[
  {
    "url": "http://mushaheedkapadia.com",
    "displayName": "Mush's Personal Site"
  } //, add as many websites as you want!
]

contacts.json should be a list of json objects that represent the contacts to text when a site goes down. A contact object has the following properties:

field	description	required
name	name of the contact	yes
phoneNumber	phone number of the contact	yes

Example contact.json file:

[
  {
    "name": "Mush",
    "phoneNumber": "+12223334444"  // use a number that can receive text messages...
  }
]

You can store these files anywhere you feel comfortable, just be sure to update the path to the .json file in config/contacts.js and config/websites.js. You can also alternatively just overwrite these two files with your .json` lists.

twilio

You'll have to set up your own twilio account since this application is a SaaS product. Once you've set up a twilio account, funded it, and gotten your authoization tokens and what not, you'll want to export these values via your environment variables.

My personal favorite is to create shell script that exports these values. Guilty Spark expects the following environment variables:

ENV VAR	description
TWILIO_ACCOUNT_SID	the account sid given to you by twilio
TWILIO_AUTH_TOKEN	the auth token of your twilio account
TWILIO_PHONE_NUMBER	the twilio phone number to send texts from

cron

It took me quite a while to get this setup with cron. I'm not super familiar with it, and you'll probably have an easier time getting it going by following some guides or fumbling your way through it.

Here's an abstracted version of the command:

1 */1 * * * [user] ([cd path/to/guilty-spark] && [. path/to/guilty-spark-env] && [path/to/node] guilty.js [>> /var/log/guilty-spark/guilty.spark.log])

NOTE: user is only valid if you throw this in /etc/crontab instead of crontab -e.

Here are some things I ran into while getting this to work:

• cron and nvm do not play well together naturally

So I'm not sure if this is because I installed nvm retardedly or something, but I was hitting an issue where the app was crashing via cron, but not when I would run the app manually. Turns out that when cron ran the job, it was setting it with minimal environment variables, and so node wasn't being set. I overcame this by running which node manually on my user and then dropping that in place for node in the cron command. Not sure

• crontab -e vs /etc/crontab

crontab -e is for your user (I'm on DigitalOcean running everythin as root #yolo) and /etc/crontab is system-wide. The fun part of /etc/crontab is that you can specify a user with whos permissions you want to run the command.

I had a hard time getting it to work, so yolo I'm running it from /etc/crontab.

• log file

Don't know why this is happening, but my log file keeps "truncating" and overwriting itself everytime the app runs. I have to look into this still, but I'm all ears if anyone has any feedback.