CASE 76 · MURMUR · 2025
Self-hosted observability that actually saves money.
A climate analytics company had been on a fully-managed observability vendor at $52k/mo. They had the engineering capacity to operate their own but had assumed it would be more expensive once you priced in the time. We built a self-hosted Grafana + Tempo + Loki stack on EKS and brought the all-in cost (including operational time) under $14k/mo.
Climate analytics
PLATFORM
2025
RESULTS
What changed, by the numbers.
ALL-IN COST
−73%
RETENTION
90d
OPERATIONAL TIME
~3h / WEEK
QUERY LATENCY (p95)
< 1.5s
HOW IT WENT
The case against self-hosting had been "engineering time costs more than the SaaS bill." We modelled it carefully: even pricing engineering time at $200/hour, the SaaS premium covered roughly 12 engineer-hours a week, which our usage model suggested would not be the steady-state operating cost.
EKS hosts the open-source Grafana stack; S3 backs Tempo (traces) and Loki (logs); AMP handles metrics for managed Prometheus. OpenTelemetry instrumentation in services replaced the vendor-specific SDK, with the bonus that it’s now portable. Backstage queries surface service-specific dashboards.
Operational time landed at about three hours a week, well under our modelled twelve. Query latency at p95 matched the previous vendor. Retention went up to 90 days because S3 is cheap. The annual saving funds two senior platform engineering hires.
RELATED · SAME DOMAIN
Other engagements in this space.
READY WHEN YOU ARE
Let's get your AWS bill (and architecture) in order.
The discovery call is free. You walk away with at least one concrete idea — even if we never work together.