Building Cerebro:
next-gen observability platform at Royal Ahold Delhaize

Evil8 partnered with Ahold Delhaize to develop a next-generation observability (o11y) platform for their TechNL IT organization, supporting the brands Albert Heijn, Etos, and Gall&Gall, comprising over 1,500 engineers. The o11y team is part of the Engineering Enablement Platform (EEP) department, which includes 150 people.

Our mission included bootstrapping a new platform, facilitating key migrations, and building a self-sustaining team for future development. Evil8 provided two interim roles, Engineering Manager and Staff Engineer, to address the following challenges:

  • Low maturity in observability practices and tooling
  • High costs due to multiple licenses and expensive infrastructure
  • Tool sprawl with various monitoring solutions, including Nagios, Elastic, ADX, Prometheus, Dynatrace, and others
  • Low cohesion from lack of standardization and shared knowledge

We delivered the platform’s MVP within 3 months, achieving 80% team adoption in under a year. This rapid adoption enabled other teams to build novel solutions on top of the platform. We restructured the o11y team and established an effective work culture through strategic hiring, coaching, TDD, continuous deployment, and high-quality engineering.

Technologically, we self-hosted the LGTM stack (Loki, Grafana, Tempo, and Mimir), with Prometheus already widely in use. Grafana facilitated connections to existing data sources. High adoption resulted from designing the platform in an as-a-service model, allowing teams to use it independently of the platform team. To support this model, we offered standardized APIs (Grafana, Prometheus, OpenTelemetry, Syslog, etc.) as integration points. We streamlined onboarding with existing IAM, prioritized comprehensive documentation, and provided first-class support. The platform’s multi-tenant design promotes efficient reuse of dashboards and alerts.

We collaborated closely with IT leadership and other platform teams focused on Kubernetes, Azure, GitHub, and more.