Grafana Mimir vs Alopex Skulk

What Grafana Mimir is

Grafana Mimir is an open source project that provides horizontally scalable, highly available, multi-tenant long-term storage for Prometheus and OpenTelemetry metrics.

How it works

Components: microservices such as the distributor, ingester, querier, query frontend, store gateway and compactor, all compiled into one binary.
Ingest: the preferred architecture puts Apache Kafka in the middle to decouple writes from reads. The classic architecture has ingesters record writes on local disk.
Long-term storage: per-tenant blocks based on Prometheus TSDB are stored in object storage.

Main features

Multi-tenancy keyed by a tenant ID on every request.
Parallelized queries that hold up under high cardinality.
Recording and alerting rules.

Limits

The preferred Grafana Mimir architecture requires Apache Kafka or a Kafka-compatible system, which is a separate product. Authentication and authorization are left to an external reverse proxy. The license is AGPL.

How it differs from Alopex Skulk

Both store Prometheus metrics and query them with PromQL.

Alopex Skulk is a time-series engine you can embed in your application. Grafana Mimir is strong at pooling many teams’ metrics into one large platform; Alopex Skulk starts where metrics are produced, storing and reading them inside the app or device, and grows to a server and a cluster on the same engine.

Before you build a central metrics platform, store metrics inside your app and read them with PromQL. Alopex Skulk shows how to start.

References

Grafana Mimir documentation, Grafana Mimir architecture, About ingest storage architecture, Grafana Mimir (GitHub)