To help monitor network health.
Exit nodes periodically execute post-routing-table.sh, hitting an API endpoint at https://peoplesopen.herokuapp.com/api/v0/nodes. This relays information about the number of active routes, active gateways, and the full contents of the exit node's routing table. If an exit node hasn't checked in in more than 2 minutes, it is assumed to be down.
Uses memcache/memjs, and mongo db. Deployed to heroku.
Leaves much room for improvement ;)
Step-by-step guide.
sudo apt-get install memcached mongodb
Open a terminal, install git and clone this repository using:
cd [your projects directory]
git clone https://github.com/sudomesh/monitor
cd monitor
This should create a new directory called monitor in your projects directory.
npm install
cp dev.env .env
memcached -v
Create a folder in monitor called data where mongo writes all of its database files, then start up the mongodb server:
mkdir data
mongod --dbpath=data
node tools/setup-db.js
npm start
Check whether your app launched properly by going to http://localhost:3000 in your favorite browser.
By default, you should see a notification that all exit nodes are down. To simulate exit nodes phoning home about their routing tables, open a commandline in the monitor directory and execute:
node tools/simulate-activity.js
This script sets a key-value pair on memcache in such a way that the monitor page gets all like "hey, this exit node is alive!"
After running this, reload your local monitor page and the page should indicate that the exit monitor is up. After 2 minutes, this is no longer the case, because the activity metrics expire.
If you ever need to wipe your mongo database of simulated activity, kill mongod with ctrl+c, make a fresh data directory, then restart mongod:
rm -r data
mkdir data
mongod --dbpath=data
To run tests, first install jasmine using npm install jasmine -g. Then, you can run the tests using npm test.
PRs (and additional commits against them) will always trigger a Travis build. Successful Travis builds will spin up a dev deployment on Heroku automatically if the branch is in the repo, and dev deployments can be triggered manually by maintainers if the PR is from a fork.
Merges into master and commits directly against master will both trigger Travis builds. Successful builds are then deployed to production.
As PRs from forks do not automatically deploy review applications, new contributors may wish to deploy their own Heroku instances. You won't be able to see your logs, but dev (free) application is otherwise sufficient for this purpose.
Make sure to enable the memcachier addon.
heroku addons:create memcachier:dev
If you you are not familiar with heroku, please see https://heroku.com.
To install the heroku cli, see https://devcenter.heroku.com/articles/heroku-cli. To check whether you installed the heroku cli correctly, run heroku --version. Expected results is something like:
$ heroku --version
heroku-cli/6.11.17 (linux-x64) node-v7.10.0
To push to current production server, get push access and add the following to .git/config:
[remote "heroku"]
url = https://git.heroku.com/peoplesopen.git
fetch = +refs/heads/*:refs/remotes/heroku/*
We are happy to receive bug reports, fixes, documentation enhancements, and other improvements.
Please report bugs via the github issue tracker.