Same signals. A structurally different bill.
Datadog's rate card is built around what you send it: host counts, custom metrics, log volume ingested. Randoli's isn't, because processing happens locally and nothing about our pricing depends on ingestion at all. That's not a discount. It's a different architecture.
"We cut our Datadog bill by 60% after moving to Randoli, without losing the visibility our platform team relies on."
Platform Engineering Lead, Canadian fintech
Where it actually differs
Ingestion pricing
$0 ingestion. Telemetry is processed locally in your cluster. No billing meter for log bytes, trace spans, or custom metric volume, at any scale.
Per-GB log ingestion plus per-host infrastructure. Two independent meters running simultaneously: host count and data volume, both growing as workloads scale.
Custom metric cardinality
Custom metrics included in the flat per-host rate. Adding Kubernetes labels or application counters doesn't add to your bill. Cardinality is your engineering choice, not a cost lever.
Custom metrics billed per thousand data points per month. High-cardinality Kubernetes workloads routinely push this cost well above the base host fee, often the biggest surprise on the bill.
Data ownership
Raw logs and traces never leave your cluster. Local processing keeps sensitive telemetry inside your environment, under your retention policies and access controls.
All telemetry egresses to Datadog's cloud. Raw logs and traces are stored in their infrastructure under their retention schedule. Compliance requirements need per-plan verification.
APM & trace indexing
Distributed tracing included in the per-host rate. No per-span indexing fee on top of infrastructure cost. Deeper service instrumentation doesn't change your bill.
APM billed per host, plus indexed trace spans billed separately per million spans retained. Growing service depth increases both meters independently.
Log monitoring
Logs stream through the same real-time pipeline as everything else. No separate bill for log volume, at any scale.
Log ingestion and indexing are billed separately from infrastructure, with processing delay before logs are queryable.
The numbers
Volume assumption: ~300 GB logs/day and ~50 M trace spans/day per 100 hosts, typical for a mid-size platform team. Randoli's per-host rate is $0.04/host/hour; Datadog estimates are based on published infrastructure, log, and APM rate cards at standard pricing.
| Scenario | Datadog (est.) | Randoli |
|---|---|---|
| ~$3,400/mo | ~$1,440/mo | |
Randoli: $0.04/host/hr × 50 hosts × ~730 hrs/mo ≈ ~$1,440/mo Datadog (est.): illustrative, based on that vendor's published rate card at the volume in this scenario — see the assumption note above. | ||
| ~$13,600/mo | ~$5,760/mo | |
Randoli: $0.04/host/hr × 200 hosts × ~730 hrs/mo ≈ ~$5,760/mo Datadog (est.): illustrative, based on that vendor's published rate card at the volume in this scenario — see the assumption note above. | ||
| ~$19,000+/mo | ~$5,760/mo | |
Randoli: $0.04/host/hr × 200 hosts × ~730 hrs/mo ≈ ~$5,760/mo Datadog (est.): illustrative, based on that vendor's published rate card at the volume in this scenario — see the assumption note above. | ||
| Quote-based, six figures annually | ~$14,400/mo (volume discounts available) | |
Randoli: $0.04/host/hr × 500 hosts × ~730 hrs/mo ≈ ~$14,400/mo Datadog (est.): illustrative, based on that vendor's published rate card at the volume in this scenario — see the assumption note above. | ||
Illustrative estimates based on published rate cards and typical usage patterns, not a quote. Confirm current numbers before publishing.
Why teams switch
Datadog is genuinely broad. Coverage across infrastructure, APM, logs, RUM, and security in one console is a real reason it became the default for platform teams. The friction shows up later, when host count, custom metric cardinality, and log volume all grow at once and the bill grows faster than any one of them did alone.
Teams that switch usually aren't unhappy with what Datadog shows them. They're unhappy that the cost of showing it keeps changing shape. Moving to a model where ingestion is structurally free means the bill is predictable again, and that predictability doesn't require cutting observability coverage to achieve it.