Your data never leaves your environment.
Most observability platforms centralize your logs, traces, and metrics in someone else's cloud before you can query them.
Randoli splits the job in two: a data plane that stays in your environment, and a control plane that ingests only signals.
Why this makes $0 ingestion possible.
Ingestion pricing exists because vendors have to pay to receive and store your data. When processing happens in your environment instead, there's no ingestion step on Randoli's side to bill for, the cost structure disappears along with the step that caused it.
Everything that touches raw telemetry runs where your workloads run.
The Randoli agent, an OpenTelemetry collector, and the storage backends behind them (Prometheus, Tempo, and Loki by default, or your own OTel-compatible backends) all run inside your own cluster. Local processing means logs and traces never leave the environment they were generated in.
Randoli's hosted side sees signals, not your telemetry.
What reaches Randoli's control plane is correlated signals and metadata, not the raw logs, traces, and metrics behind them. That's the basis for the data-sovereignty claim used across the rest of the site: retention stays your own decision, on your own infrastructure, at no extra cost.
Standard instrumentation, not a proprietary agent.
Randoli ships an opinionated OpenTelemetry setup so teams can start fast, but it's standard OTel underneath: no proprietary format to adopt and no lock-in to unwind later.
See the full setup and language coverage on OpenTelemetry.