STORM

Enter your email address to get free access to STORM

No Credit Card Details Required.

x

What we've learnt about deployment

Written By

Tim Dunton

Published on

August 20, 2019

Category

Evaluation

As some of you may already know, we've had the deployment feature in development for over six months. For those who don't know, deployment allow you to deploy your code directly from a particular branch and repository from either GitHub or BitBucket. This means your code can be on the server in seconds instead of having to use an FTP or SSH client to upload your files.

When we first built this feature we realised it would be a killer feature for STORM. The reason being that all the existing methods for doing this were very complicated and either needed an SSH key, whitelisted IP addresses or a complicated command line toolkit. Like with lots of features in STORM, such as Let's Encrypt for example, this feature is like an iceberg. There are only a few configurable options available to the user, but in the background there are thousands of lines of code to make the process as easy as possible.

In fact, since we've launched deployment it has received more feedback than any other feature inside STORM. Lots of our STORM users have asked us to extend the functionality and we certainly intend on doing that. 

Three problems

Since launching the deployment feature, we have hit three major problems. The first being the complicated user interface. The idea behind using Git and deployment is actually quite complicated, and trying to simplify this is actually more difficult than we realised. In the last few weeks, we have launched an improved version which should make understanding this process a lot easier. Saying that, you need to have a good understanding of how Git works before you can use it. We have some tutorials and videos in the works to help agencies who have not yet got to grow with Git deployment.

The second problem we hit is that GitHub and BitBucket regularly change their API with either very little notice or no notice at all. This means that deployments can easily fail, and no one is wiser. We have improved this by adding in email notifications and warnings should things go awry. 

Thirdly, every client seems to have a different deployment process. Some want to do the entire site while others just want to do a directory. Some users update the server via FTP, causing a conflict while others do everything through Git. We have not completely managed to resolve this use but are looking to add some additional functionality to make the process easier.

For now, the feature is still in beta testing. We want to get this revised version out of beta as soon as possible so we can move onto expanding some of the features. However, our current priority is to have the existing features working thoroughly before we add any additional functionality.

Please do keep sharing us your feedback! 

Subscribe via email