Pre- and post-deploy approvals
Use manual intervention steps in Octopus at the start or end of your deployment to have people from different teams verify and approve the deployment, either before it begins or after it succeeds.
Verification and canary deployments
Use manual inteventions in the middle of a deployment to have a human check that the system is working as expected, before proceeding with the deployment. If something is wrong, they can enter a reason and fail the deployment.
Things that can't be automated
Some things can be time consuming to automate, and maybe you just haven't gotten around to it, or convinced people that it's safe to automate (we're looking at you, DBA's!). Use manual intervention steps to pause a deployment at any stage and ask a human to do something - like running database change scripts, or dealing with a legacy system - before proceeding.
Imagine you are halfway through deploying a website to a dozen production servers, when something goes wrong - maybe a file was locked or OS updates caused the machine to reboot. Instead of the whole deployment failing, you can have Octopus pause the deployment and ask for your guidance. You can fix the issue and try again, or skip that machine. Once production is back online you can review the deployment log to figure out why it failed, and prevent it next time.
Block releases from progressing
Sometimes the deployment is perfect, but you later discover a bug with the app itself - maybe a critical bug or issue that means that release should never be allowed to progress to production. You can mark the release as "blocked" in Octopus to ensure it isn't accidentally promoted.
You might also like...
@OctopusDeploy Really liking the new GUI in @OctopusDeploy 4.0, both good looking and fast. Great work!— Johan Boström (@zarx) November, 21 2017
I’d stand by @OctopusDeploy every day of the week. Honestly couldn’t imagine our CD working as smoothly without it.— Ben Macpherson (@BenjiMac93) November 18, 2017