The challenge
One cloud, many ways of writing code
Hydro’s DevOps team is spread across many countries and business units. Each team had built up best practices that worked well for them, so infrastructure code naturally looked a little different from team to team.
With so many teams, reviewing a change took time. Before approving it, someone had to work out what it did and how it would affect other repositories and environments.
Writing a change took time too. Engineers had to look up Hydro’s best practices in Confluence, the existing repositories, and Azure’s own best practices and policies, across many different systems.
“The agents handle day-to-day work, giving our engineers space to focus on improving our systems.”
Istvan Papp
Cloud Architect, Hydro
Why Cloudgeni
Best practice, built into every pull request
Hydro had already written down its best practices in Confluence. Cloudgeni reads that context every time it writes a pull request, together with the repositories and Azure policies, so engineers no longer have to look it all up themselves.
Because the agents follow the same standard every time, code is written the same way across every team and country. Changes are easier to review, easier to repeat, and the engineers pick up the best practices simply by seeing them used.
How it works
From context to a standard pull request
01
Read the context
Cloudgeni pulls Hydro’s best practices from Confluence, plus the repositories and Azure policies.
02
Write it the standard way
The agent writes the infrastructure change as code and opens a pull request.
03
Review and ship
Every change follows the same standard, so it is quick to review and merge.
Drift, fixed overnight
From a monthly check to every Monday, automatically
Hydro used to check for drift between its cloud and its code once a month, and many changes made by hand in the portal were left behind. Now Cloudgeni’s DriftGuard checks every Monday and fixes drift between the cloud and the code automatically.
It runs overnight. All the routine work is done before the team arrives at 9am, so the CI/CD pipelines are not overloaded when everyone starts making changes.
Results
Time, money and quality
- Time
- Hours no longer spent looking up context in Confluence and Azure policies, and far less time spent on drift.
- Money
- $200K saved, because Hydro did not need to hire more DevOps engineers.
- Quality
- Code quality went up, and more of the cloud now lives in code instead of manual changes in the portal.
