Maintenance to Momentum: Why Celonis Made the Switch from Self-Hosted Backstage to Roadie
Published on July 29th, 2025 by Jian ReisAbout Celonis
Celonis is the global leader in Process Mining, helping organizations identify inefficiencies and optimize operations across finance, supply chain, customer service, and beyond. With over 3,000 employees, more than 20 offices worldwide and 1,500 enterprise customers, Celonis powers process excellence for more than 30% of the Fortune Global 500.
Celonis’ OSS Backstage Era
Internally, Celonis brings the same performance-driven mindset to its engineering operations, investing in tools that help teams move fast, operate safely, and scale with confidence. That mindset is what led them to adopt open source Backstage in 2022: a flexible, extensible platform that offered a centralized view of their software ecosystem, clear ownership, and a framework for codifying engineering standards into daily workflows.
The initial goal was to build a unified internal developer portal that could drive consistency and visibility across teams. And in the early days, Backstage certainly delivered, improving discoverability and reducing reliance on tribal knowledge.
According to Andreas Bayer, VP System Engineering at Celonis:
A few of the early wins were consolidating documentation scattered across GitHub, Confluence, and Google Docs into one place, and improving discoverability. As the company grew, Backstage also helped address service ownership: previously, everyone knew which team owned what, but that model broke down as we scaled. Having a central catalog helped resolve that.
But over time, the limitations of maintaining Backstage internally began to surface.
A team of 3 FTEs is not able to deliver a sufficiently useful portal based on OSS Backstage. There are obvious pieces like Tech Insights, but also many little things that add up to a much better user experience - things that would take weeks to build and are hard to prioritize over bigger features. The maintenance overhead is just too high.
Keeping up with backend upgrades was one of the biggest drags on velocity. As Andreas described it, it became a constant trade-off: invest time in upgrades or in feature development. Every hour spent keeping Backstage alive was an hour not spent improving the developer experience.
It was always a choice between building new features and falling behind on maintenance, or staying up to date but delivering little value to users. That’s not sustainable.
According to data collected for Roadie’s 2025 State of Backstage Report, 70 percent of companies that are ‘very happy’ with Backstage dedicate at least 3 full-time engineers to maintaining it. For teams that can’t justify such a level of investment, maintaining a production-grade experience becomes a game of trade-offs.
The tipping point came gradually. Over time, the team noticed that the very features sitting on their internal backlog were already being shipped by other platforms like Roadie: Scorecards, catalog UX improvements, integrations - they were constantly being outpaced.
We saw vendors shipping features we’d always wanted to build but never got to. The final nudge came when we wanted to double down on scorecards and readiness checks. Building and wiring up the Tech Insights stack ourselves was so time-consuming - then we looked at Roadie and realized, wait, it’s already there. No engineering lift required.
Making the Case for Roadie
That’s when Celonis began to seriously evaluate Roadie. Unlike a from-scratch rebuild or a major migration project, Roadie offered a natural next step for an existing Backstage adopter: a fully-managed Backstage platform that could give the Celonis team everything they needed, without the overhead.
Roadie is Backstage, just with the batteries already included. The compatibility was crucial: Celonis could reuse their existing catalog-info.yaml files, extensions, and the knowledge and patterns they had already established. They got continuity, without the cost of ongoing maintenance.
One of the biggest draws was that Roadie is deeply involved in the Backstage ecosystem: we’d seen Roadie’s contributions to ArgoCD plugins and other parts of the platform. That gave us a lot of confidence in their technical approach and commitment.
Beyond compatibility, Roadie gave Celonis:
- A developer portal that was polished, fast, and production-ready; no weeks-long sprints just to improve search, navigation, or quality of life features
- A fully integrated Tech Insights platform, letting them build scorecards and enforce standards from day one
- Powerful customization options like agent-based ingestion and the Fragments API to support more advanced use cases
- Freedom to focus on value, not infrastructure - Roadie took on the burden of plugin upkeep, upgrade stability, performance tuning, and support.
All the usual pain points - plugins breaking, authentication weirdness, long CI cycles, slow page loads, they just go away. It’s not just a hosted Backstage - it’s a managed solution that lets us focus on what actually matters to our engineers.
Migration: From Evaluation to Execution
Celonis engaged Roadie through a Proof of Value (PoV) evaluation, giving them the chance to test capabilities, validate integrations, and work closely with Roadie’s support and solutions engineering teams.
Following a successful PoV, the decision was made to fully migrate off their self-hosted instance. Thanks to the shared Backstage foundation, much of the transition was seamless. Existing entity definitions, scorecards, and metadata were immediately portable. Within two weeks, the production migration was complete.
Hooking up our existing catalog files just worked. But the real wow moment came when the team realized how easy it was to write new scorecard checks. People who had struggled with the old system were like, ‘wait, that’s it? No dev containers? No plugin juggling? It just works!’
Most of the effort went into trimming internal customizations: choosing what to keep, what to drop, and what to rethink now that Roadie offered native alternatives. As Andreas puts it, the migration was as much a cleanup opportunity as it was a platform shift:
It helped us revisit old decisions and remove customizations that weren’t as useful as we originally thought. The technical migration was easy - most of the work was organizational.
What’s Next
Now that the migration is complete, Celonis is focused on scaling the value of their developer portal.
The priorities include:
- Expanding Tech Insights usage, including custom filters and compliance labels unique to their requirements
- Lifecycle gating to ensure services meet quality thresholds before going live
- Sorting, filtering, and custom columns to help platform and security teams prioritize risk
- Future plans for automated remediations, Slack notifications, and Scaffolder-driven workflows.
Celonis is now also rebuilding their scaffolder strategy with the goal of making service creation fully self-service, from repo setup to monitoring and documentation. Longer term, they’re exploring how templates can streamline squad registration and ownership updates in line with their org model.
Advice to Other Teams
When asked if Celonis would recommend other teams who are self-hosting Backstage to consider switching to Roadie, Andreas was emphatic:
It’s definitely worth switching. With Roadie, you get to spend your time thinking about what’s valuable for your users - not worrying about Backstage internals or upgrade mechanics.
Thinking of switching to managed Backstage? See how Roadie can help: book a demo, contact us at sales@roadie.io, or check out our Backstage comparison guide.