Email iconarrow-down-circleGroup 8Path 3arrow-rightGroup 4Combined Shapearrow-rightGroup 4Combined ShapeUntitled 2Untitled 2Path 3ozFill 166crosscupcake-icondribbble iconGroupPage 1GitHamburgerPage 1Page 1LinkedInOval 1Page 1Email iconphone iconPodcast ctaPodcast ctaPodcastpushpinblog icon copy 2 + Bitmap Copy 2Fill 1medal copy 3Group 7twitter icontwitter iconPage 1

Quick links

Recap

After our second hack day we had a fully functional hUVD (heads up visual display). I think the whole team would agree that it’s a great tool and we all find it really useful to see all of the most important info, in one place. In true Agile style, we’ve a backlog of features we would like to add, but we decided that the most important addition would be to focus on our ability to support our customers.

How we currently monitor our client applications

We host two large scale web applications; Limpid Markets and Havells Sylvania, and have worked closely with both clients on an SLA. Due the importance of the stability of the applications, we’re using Amazon’s CloudWatch to monitor the servers and Winston to monitor Node application errors. Notifications from these services are sent to us using Amazon’s Simple Notification Service and in the past we’ve handled these through email notifications sent from the service to a series of mailboxes (such as ‘alerts@uvd.co.uk’).

Hooking into CloudWatch

This is all brilliant ‘out the box’ stuff from AWS  but as we’re a small team there are lots of potential issues with relying on one or two people to monitor errors via email (holidays, lunch breaks, illness, forgetfulness, resignations and so on) and let’s face it, being the support gatekeeper is not that glamorous. To solve this problem we agreed that it would be beneficial to make support more transparent and a team responsibility, so following discussions, we thought it would be great to add an alert system to the hUVD which hooks into our CloudWatch system.

Dardan updated the Node application to listen for the notifications coming in from the Simple Notification Service and then sending those notifications to the hUVD via websockets. The application is also used as a server proxy in order for the hUVD dashboard and admin frontend projects to communicate with each other with socket.io. This means they can share error messages from AWS, and send ‘dismiss’ commands to remove the notification popup.

James implemented a new AngularJS module into the existing application which listened for errors provided by the AWS SNS. The hUVD then alerts us UVD’ers that an error had occurred with unmistakable urgency.

The end result

Thoughts from the team


James

James

Frontend Developer

“The hUVD really allowed me to get to grips with Unit Testing in JavaScript as well as helping to drive home the Behaviour Driven Development principles. It was great to ease myself into the process by working on an internal project.”


Once the backend was functioning our new Digital Designer, Alvaro set to work on how it looks and James kindly ensured that it has THE most annoying warning sound, ever.


Alvaro

Alvaro

Digital Designer

“Getting to design the notification message of the hUVD may seem a small step for anyone else, but for me, as I needed to work with SASS code for the first time, it was very interesting. I can’t wait to get stuck into designing and writing SASS for live client work!”


Share: