Delivery Automation

Deployments should be boring. At Norconsult Astando I took them from about 45 minutes to about 5; at Plejmo, from an hour to under two.

GitHub Actions, Azure Pipelines, Octopus Deploy or whatever the team already has. It also brings a few practical advantages:

• It saves time: deployments become a non‑event instead of a time sink
• It’s reliable: the same steps run the same way every time
• It’s person‑independent: anyone can trigger a release, or it can run fully automatically

Below are two examples of how I’ve helped teams move from manual deployments to predictable, automated delivery.

Norconsult Astando AB

Isy Road had an all manual deployment process. The deployment documentation was spread over several documents and was not up to date. I helped the client setup automated deployment for Isy Road as a pilot project. Octopus Deploy was used and benefits included:

• Reduced deployment time from ~45 minutes to ~5 minutes per application instance - 15 at the time.
• Automated database deployment.
• More confidence in the deployment process.
• Virtually no deployment related errors.

The success of the pilot project lead to setting up automated deploy for their products Isy Map and Isy Case as well.

Plejmo

Plejmo had an all manual deployment process that was not very difficult but stole a lot of time from the small development team ‐ time that should be used to develop the product. Using Octopus Deploy we were able to:

• Reduce deployment time from ~1 hour to under 2 minutes.
• Automate database deployment.
• Virtually eliminate deployment related errors.

The deployment process became so quick and painless that we could (and often did) deploy several times a day ‐ even during peak hours.