Introduction
When Google announced it was deprecating its Container Registry (GCR), hundreds of companies were forced into a time-sensitive migration. For Outreach, a sales execution platform with over 100 microservices, this wasn't just an inconvenience—it was a high-stakes project with a non-negotiable deadline. Failure to migrate meant their teams would no longer be able to deploy code.
The challenge was compounded by their complex, template-driven microservice architecture. Egor Kalinichev, Engineering Manager for the Dev Tooling team, was responsible for ensuring every single service was updated. But with so many moving parts, how could he track the progress of such a critical, company-wide initiative without resorting to spreadsheets, Slack harassment, and "scary stories"?
By leveraging a custom Cortex Scorecard, Egor and his team were able to:
Gain 100% visibility into a complex, 100+ service migration, tracking real-time progress against a hard deadline.
Proactively identify and support lagging teams, ensuring no service was left behind.
Eliminate the stress and chaos typically associated with time-sensitive migrations, allowing teams to maintain their regular development pace.
Provide leadership with confident, data-backed updates, replacing anxiety with accurate progress reports.
The challenge: A ticking clock and a black box
In mid-2024, the clock started ticking. Google's decision to retire GCR meant Outreach had to move its container registry to Amazon's Elastic Container Registry (ECR) to co-locate it with the rest of their AWS workloads. The migration itself was clear, but the path to get there was anything but.
Outreach manages its hundreds of microservices with an internal open-source tool called stencil, which templates everything from logging to the CI/CD pipeline. To migrate to ECR, the Dev Tooling team had to update the core stencil modules that defined how services were built and pushed to a registry.
The real problem wasn't updating the templates; it was tracking the adoption of those updates across more than 100 services, owned by dozens of different teams. Without a centralized system of record, Egor's only option was a manual, high-friction campaign.
"If we did not have anything at all, administrative methods would be our only choice. I would be writing to all the managers telling them if they didn't do what we asked, they wouldn't be able to deploy."
Egor Kalinichev, Engineering Manager, Outreach
This approach would have created a storm of confusion and anxiety. Engineering managers would be scrambling, developers would be interrupted, and no one would have a clear picture of how much work was left to do.
The solution: A Scorecard for clarity and confidence
Instead of resorting to this heavy-handed management approach, Egor turned to Cortex. He created a custom Scorecard called "ECR Readiness" that became the single source of truth for the entire migration.
Using Cortex Query Language (CQL), the Scorecard was configured to parse a stencil.log file within each service's repository. This log file contained the version data for every stencil module the service used. By checking these versions, the Scorecard could automatically determine if a service had been updated for ECR compatibility.
The result was a real-time, "bird's-eye view" of the entire engineering organization, neatly broken down by team and service.
"The scorecard was able to give me a report of who already has the latest version and will survive the shutdown of the container registry, and more importantly, who will not. I was able to dial down on who were the biggest offenders, who were slow to upgrade, and who I needed to help."
Egor Kalinichev, Engineering Manager, Outreach
With this data in hand, Egor could be surgical in his approach. He knew exactly which teams to contact, which services needed attention, and could even identify recently inherited services where the new owners might be unaware of the migration requirements.
The impact: A migration without the mayhem
The most significant outcome of using Cortex wasn't speed—it was the reduction of overall stress across the engineering org. In a scenario that could have easily devolved into a high-pressure, "all hands on deck" fire drill, the Outreach team was able to proceed with calm and confidence.
Because they had perfect visibility into their progress, they could pace the migration over two months without derailing other projects. Leadership, in turn, received clear, data-driven updates that replaced uncertainty with confidence.
For a dev tooling team supporting an organization of 300 engineers, the value is clear. When asked about the ROI of Cortex, Egor compared it to the cost of another engineer.
"If I had to choose between hiring one more person or purchasing Cortex, it would be an easy choice: Cortex definitely makes a bigger difference."
Egor Kalinichev, Engineering Manager, Outreach
Automating insights for proactive improvements
The success of the ECR migration has inspired Outreach to deepen its integration with Cortex. While the CQL-powered Scorecard was powerful, Egor's team is now working to make this kind of data even more accessible.
Their next initiative is to use the Cortex API to have their CI pipeline automatically push stencil module versions as custom data to the Service Catalog. This will make building future Scorecards significantly easier and will enable them to proactively nudge teams about updates more frequently, further improving the overall health and security of their codebase.
"This is the angle we look at Cortex from right now. We can utilize it so much more and improve the developer experience so much more. So that's what we're going for."
Egor Kalinichev, Engineering Manager, Outreach
Ready to manage your next migration with confidence? See how Cortex can help your team achieve similar results. Book a demo to learn more.
