Pricing model deep-dive
Per-GB Ingest RUM Pricing Models 2026
A handful of vendors meter RUM cost by the volume of telemetry data ingested, in GB per month, alongside APM, logs, and infrastructure metrics. This shared-meter model is cheap for low-volume RUM but couples RUM cost to total observability spend.
Direct answer
How much does per-GB RUM ingest cost compared to per-session?
A typical RUM workload sends ~40MB per 1,000 sessions. At New Relic's $0.40/GB rate, that's $0.016 per 1,000 sessions, dramatically cheaper than per-session vendors. But ingest is shared across products (APM, logs, infra), so the total observability bill depends on the broader stack, not RUM alone.
Vendors using per-GB ingest pricing
- New Relic Browser — $0.40-$0.60/GB shared with APM, logs, infra.
- Elastic Browser RUM — ingest tied to Elastic Cloud cluster size (~$95/mo+ cluster cost).
- Grafana Cloud (Grafana Faro backend) — per-GB ingest on Loki/Mimir/Tempo.
The shared-meter trap
The defining feature of per-GB ingest pricing is that RUM data and other observability data (APM traces, infrastructure metrics, application logs) all count against the same monthly ingest budget. This produces two operational dynamics:
- Cheap when other ingest is small. If RUM is your only observability product, per-GB pricing is dramatically cheaper than per-session pricing at the same workload.
- Expensive when other ingest dominates. If you're running 500 hosts of New Relic APM at 1TB/month of trace ingest, adding RUM is a marginal addition. But the total observability bill is much larger than RUM alone.
Sample math at 1 million sessions/month
At ~40MB per 1,000 sessions, 1 million sessions/month generates ~40GB of pure RUM ingest. At $0.40/GB:
- RUM-only on New Relic: 40GB × $0.40 = $16/month for RUM. Plus 100GB free tier covers the entire workload, effectively $0.
- RUM added to existing 200GB/mo APM workload on New Relic: 200GB → 240GB total. Net incremental RUM cost: $16/month. But total bill is now 240GB × $0.40 = $96/month.
- RUM added to existing 2TB/mo logs workload on New Relic: 2000GB → 2040GB total. Net incremental RUM cost: $16/month. But total bill is 2040GB × $0.40 = $816/month, mostly logs, RUM is rounding error.
When per-GB ingest wins
- You already run the vendor for APM, logs, or infrastructure and want to add RUM under the same contract.
- Your RUM ingest is small relative to other observability ingest, RUM becomes a marginal cost addition.
- You have a single-vendor observability strategy and want unified billing.
- You can predict total ingest volume across all products for committed-spend annual discounts.
When per-GB ingest loses
- You only need RUM in isolation — the shared-meter model has no advantage and the per-GB rate ($0.40+) is more expensive than per-session at very large volumes.
- Your team can't accurately forecast cross-product ingest — budget surprises are common.
- You want per-product cost attribution for internal chargeback — per-GB shared meters obscure this.
Where to go next
- For per-session pricing structure, see per-session pricing models.
- For per-pageview pricing structure, see per-pageview pricing models.
- For New Relic, Elastic, Grafana per-vendor detail, see New Relic Browser, Elastic Browser RUM, Grafana Faro.
Last verified June 2026