Koz Software

My focus is self-hosted observability, especially metrics and alerting, on shared infrastructure: the Prometheus and Grafana Mimir platforms a central team runs for many others. I know those systems from the inside, including how ingesters, the ruler and the query path behave under real multi-tenant load, and I also know the organisational work that comes with them.

Alongside client work I'm building obscost, an open-source tool that works out which tenants and rules are using a shared Mimir cluster, and what that costs.

How I can help

  • A single session on a specific problem, such as a cardinality spike, a lagging ruler or an alerting setup nobody trusts.
  • A platform review of your metrics and alerting stack, ending in a written, prioritised list of changes.
  • Hands-on delivery, working in your repositories and clusters alongside your team, for a migration, a scaling push or new tenant cost reporting.

Tools I use most

Prometheus, Grafana Mimir, Alertmanager, Grafana, PromQL, Kubernetes, Go, Python, SQL, and AWS.

Get in touch

The quickest way is to book a free 30-minute call. You can also email me at admin@kozsoftware.com.