Posted on Fri 24 May 2013

Deployment is Scary - Part 2

The is the second installation of a series of posts about my deployment experiences with Enteract. Part 1 is available here.

In this post, I will focus on our use of Faraday and Upstart for some simple monitoring.

Faraday

We wanted a tool to monitor the web application and notify a member of the team should the web application be unreachable or defaced. The first requirement was relatively simple, and can be easily satisfied by using Amazon's Elastic Load Balancer with the health check gem for Rails.

The second requirement was slightly trickier, and I wrote a ruby script using Faraday that periodically

  • visits the sign in page,
  • signs in,
  • visits the home page,
  • ensures that the names of all 24 entries exist on the page, and
  • notifies the team, via text messages, should any of the assertions fail.

(On hindsight, it would have been much easier to use Capybara or JMeter for this purpose; I was using this as an excuse to learn how to use Faraday and Amazon's Simple Notification Service.)

I uploaded a gist of the relevant script with the sensitive information redacted. The only tricky part was retrieving the CSRF token and preserving cookie information between requests.

Upstart

Next, I wanted to some reasonably reliable way to ensure that the monitoring process is alive - it wouldn't be a very useful monitoring tool if the process terminates prematurely. For this we used Upstart, an event-based init daemon that supervises the process and restarts it if it terminates.

The configuration file that worked for me is as follows:

There is probably an easier to set this up that uses a shorter command.

Category: Software

Tags: ruby, rails

Jim LimGitHubLinkedInStackOverflowontoplist ⋅ Theme by Giulio Fidente.