Tldr: great spike to solve deployment problem may be a wasted effort.

Deployments of an ancient electron application need to be done in CodeDeploy to deploy the latest build. Customer hour restrictions cause this to be done only after midnight, and manually checked.

The whole team knows this is the wrong method of deployment and that there are many other operational problems with the project.

A few other senior team members get together and decide to spike out a way to use electron auto-deployment to accomplish this without using code-deploy at all.

After a shallow dive into this subject, we all get pulled aside to handle a change in another part of the software ecosystem. It happens. We leave the spike behind.

A junior-intermediate developer on the team pics the project up and gets a good spike going in a day and a half! We are all high fives and beers. This is Friday.

By Monday there is a pull request in for code review and it looks solid. Seems like it will make deployments a lot better.

Preparing the last deployment (hopefully) with CodeDeploy ever...

Marketing team members inform us that they are running an add system on the customer devices and to do it they are using Linux.

The current application being deployed is using Windows 10 (yeah, another problem).

They say they have made plans to move our application over to Linux. This means we may not be able to launch the junior devs great spike and the old deployment method may stay for the time being.

Meetings soon to find out how all of this will hash out.

End of rant. I hope I'm doing this right

Add Comment