Cut observability spend by 60% on AWS without losing production signal.
Etalon migrates enterprises from Datadog, Splunk, New Relic, and Dynatrace to self-hosted Grafana, Loki, Tempo, and Mimir with OpenTelemetry-first pipelines, Terraform-managed infrastructure, and GitOps delivery in your AWS account. Work is done by senior platform engineers in Bucharest who have shipped LGTM-class stacks in production for years, not by a bench of juniors learning the stack on your bill.
Bucharest, Romania ยท Engagements across the EU, UK, and North America
Organizations Etalon Systems has shipped with
The observability tax
Datadog and Splunk did not become expensive by accident. Pricing follows cardinality, retention, and ingest volume. The bill grows faster than headcount because every new service adds traces, metrics, and logs that hit the same commercial meters. Finance sees a line item that rivals managed data warehouses. Engineering sees dashboards that work until someone changes a label and the bill spikes.
Splunk and Datadog also centralize control. Sampling rules, routing, and retention live in vendor consoles. That is fine until you need EU residency, air-gapped regions, or an audit trail that proves telemetry never left your boundary. Then the same convenience becomes a compliance negotiation and a procurement cycle.
Most platform teams can self-host Grafana, Loki, Tempo, and Mimir. The hard part is the six months of careful migration: dual-write, backfill, cardinality discipline, SLOs that still fire, and runbooks the on-call engineer trusts at 03:00. Without that focus, the migration stalls or production regresses and the CFO sends everyone back to the renewal.
What we deliver
LGTM migration
We deploy Grafana, Loki, Tempo, and Mimir in your AWS accounts with Terraform modules, environment separation, and GitOps promotion so changes are reviewed like application code.
You get dashboards, recording rules, and alert routes ported with parity checks against the incumbent vendor so cutover is measured, not guessed.
OpenTelemetry instrumentation
We standardize traces, metrics, and logs on OpenTelemetry collectors and processors so sampling, tail sampling, and routing are yours to change without a contract amendment.
Pipelines are vendor-neutral by construction, which is how you keep portability after the consultants leave.
Platform engineering
We wire CI/CD, secrets, backups, capacity plans, and incident playbooks into the same repositories that define the stack.
Your team receives runbooks written for pager duty, not slide decks, so the platform stays operable when the engagement ends.
Why Etalon
Etalon was built by an electronics engineer who spent years in open-source observability and AWS platform work. The pattern was the same in every large account: excellent engineers trapped under a renewal that funded features they did not need while basic questions about cardinality and retention stayed unanswered. The product gap was never “Grafana versus Datadog.” It was execution risk on a multi-quarter migration with production on the line.
Romania sits inside the European engineering tradition that treats open source as infrastructure, not a hobby. Teams here ran Prometheus, Loki, and Grafana in production while US peers were still routing everything to SaaS contracts. That is not a labor arbitrage story. It is a different default about who owns the telemetry path and who signs the pager when it breaks.
Etalon takes its name from the reference standard: the artifact other measurements calibrate against. Engagements end with Terraform state, Git history, and runbooks in your organization. We do not anchor renewals to your incident volume. If we did our job, you stop needing us for day two operations.
Move observability onto AWS you control, with LGTM and OpenTelemetry, and cut spend by roughly 60% without trading away signal.
Book a 30-minute architecture reviewDirect email: hello@etalon.systems