Scale for Prometheus and Grafanas
Scale covers running Prometheus beyond a single server and keeping data for longer. Local storage has limited retention, so teams use remote write to systems such as Thanos, Cortex or Mimir, federation and sharding, and consider the wider observability stack of logs and traces.
Stage: Grafana and Scale · Part of the Prometheus and Grafana roadmap · Last reviewed 19 November 2026
What to learn in Scale
- Set storage retention and understand TSDB blocks on disk
- Configure remote_write to send data to long-term storage
- Explain the difference between federation and remote write
- Reduce load by dropping unneeded series and raising scrape intervals
- Relate metrics to logs and traces in an observability setup
Scale interview questions
How can Prometheus keep metrics for a long time?
Local TSDB retention is limited by disk, so long-term data is usually sent through remote write to a backend such as Thanos, Mimir or a managed service, which stores it cheaply and supports global queries across instances.
What is federation in Prometheus?
Federation lets one Prometheus server scrape selected series from another through its /federate endpoint. It is used to build a higher-level view from per-cluster servers, but it is not a replacement for long-term storage.
Where Scale fits in the Prometheus and Grafana roadmap
Scale belongs to the Grafana and Scale stage (advanced level). Visualise and grow.
Other units in this stage
Back to the full Prometheus and Grafana roadmap