Deploy Rails Upgrade #2872

[rsgonzal]

Before #2872 gets deployed we need to:

• Update error message in s3
• Create Rollback plan

After deployment we need to:

• Do some general QA on production
• Try Checkr
• Do activities that will trigger jobs (emails)
• Revert custom error message

[shaun-technovation]

I see that Heroku has a maintenance mode and an associated maintenance mode page. I'm learning about all of this as we go, but I think one way to go about it would be to turn on maintenance mode, do the upgrade, and then turn off maintenance mode.

Here is a link to customizing the error/maintenance pages:
https://devcenter.heroku.com/articles/error-pages#customize-pages

Also, I randomly stumbled on Review Apps:

Review apps run the code in any GitHub pull request in a complete, disposable Heroku app. Review apps each have a unique URL you can share, making them a great way to propose, test, and merge changes to your code base.

You can configure review apps to spin up automatically for each pull request, or you can create them manually from your app’s Pipeline page in the Heroku Dashboard.

I think this could be really useful!!!

[shaun-technovation]

Looks like updating the error page is pretty straight forward.

There is a Heroku configuration setting named ERROR_PAGE_URL, this specifies where the error page is in S3: //s3.amazonaws.com/technovationchallenge.org/error_pages/server_error.html

To update this in S3:

• Login into AWS
• Navigate to S3 (click Services -> Storage -> S3)
• Under Buckets, click "technovationchallenge.org", and then "error_pages"
• Then server_error.html will be there to update

We can add another Heroku setting and associating page for MAINTENANCE_PAGE_URL.

We just need the verbiage for both the error page and maintenance page, and updating/adding them shouldn't be a problem. 👍 😃

[shaun-technovation]

Regarding rolling back, the commit before the Rails upgrade, and where Production is at now, is: 1ea023448.

So, if we need to rollback, I think we can either:

• Leave our commit history unchanged and run something like: git push production 1ea023448:master, to put Production back to where it is now
• Or update our commit history, by resetting back to 1ea023448, and then that can go through the normal deployment steps (QA -> Prod)

[shaun-technovation]

Next steps for this one, I think we just need a day/time for the deployment and also a message regarding the deployment (for the error page), and then I think we'll be ready to do the deploy, and whatever comes out of that! 😅

[rsgonzal]

What about just replace the regular text with something like:

We're sorry, but the site is currently down for a scheduled update please check back soon.