Data Ingestion Pricing: Examples & Companies

What is it

Data ingestion pricing is a billing unit where customers pay per gigabyte of logs, traces, metrics or other telemetry sent into the platform, usually before retention or query costs are added.

The meter sits at the front door. A log line, span or event is counted the moment it reaches the vendor, whether anyone ever searches it or not. Datadog and Dynatrace both bill most monitoring per host but switch to per-GB metering for logs and other bulk telemetry, where host count says little about volume.

Ingestion is distinct from its two neighbours. Per-GB storage pricing charges for data at rest over time, and per-log pricing counts records rather than bytes. Vendors favour the ingest meter because it tracks their own cost: every gigabyte has to be parsed, enriched and written. Buyers find it harder to forecast because volume is set by how engineers configure logging, not by anything procurement controls. For how usage events are captured and counted upstream of a bill like this, see tracking and metering usage events.

How it works

Each gigabyte arriving at the platform is counted against a monthly allowance, and anything above it is billed at a per-GB rate. The ingest charge is rarely the only one. Most vendors add a second or third meter on the same data for indexing, retention or query.

Unit math: Monthly cost = max(0, GB ingested − included GB) × ingest rate + (downstream meter × units kept)

Worked example (Datadog Logs). A team ships 5,000 GB of application logs a month but indexes only a 20% slice, about 150M events, at 15-day retention. Ingestion is 5,000 × $0.10 = $500. Indexing is 150 × $1.70 = $255. The bill is about $755/month, and a third of it comes from the meter the “$0.10/GB” headline leaves out. Indexing everything instead of a filtered slice would push the second line well past the first.

Companies using this

Eight companies in the corpus meter ingested gigabytes. Seven sell infrastructure, application or security observability, and Arize AI applies the same meter to LLM traces. Compare them on how much of the bill sits after ingest, not on ingest rate alone.

Patterns observed

• The free allowance stays put while the paid rate moves. New Relic’s per-GB ingest price has gone up every one to two years since its 2020 relaunch at $0.25/GB, rising 60% in total, while the headline free tier and the $99/user rate stayed nominally flat. Grafana’s free-tier allowance has not changed since at least January 2022, even though the Pro plan around it was rebuilt at least three times: a flat $49/month bundle, then per-active-user billing, then a $29/month bundle, then today’s per-product metering. The free allowance acts as a fixed anchor for new users, and pricing changes land on customers who have already scaled past it.
• Price is set by how hot the data is, not only how much there is. Sumo Logic burns 25 credits per GB for Continuous ingest, 12 for Frequent and 5 for Infrequent, so the same gigabyte can cost five times as much depending on how quickly you need it searchable. New Relic’s Data Plus tier works the same way: it buys up to 90-day retention against at least 8 days on Original data.
• Discounts come from changing what you send. Splunk weights eligible Cisco network telemetry at 0.5x of the normal ingest rate, a lever created by Cisco’s ~$28B acquisition. Grafana sells Adaptive Logs, which drops commonly ingested but unused log patterns and is advertised as cutting log-ingest cost “by up to 50%”. In both cases the vendor keeps the per-GB rate and lowers the bill by changing which gigabytes are counted.

Counterexamples & variants

“$0 ingest” that still costs money. In March 2024 Sumo Logic became the first log-analytics vendor to advertise $0-per-GB ingest, replacing a $3.00/GB log-analytics rate. Its own Cloud Flex Credit worked example shows a routine 10 GB/day workload burning 91,250 credits a year on Continuous GB Ingest alone, about $13,688 at the Essentials list rate of $0.15/credit. The meter was renamed and moved into a prepaid credit pool. It was not removed.

Leaving the ingest meter altogether. Splunk’s legacy licence bills per GB indexed per day. Third-party analysts put list rates at $150–$200/GB/day, and a 313-point Hacker News thread about replacing Splunk with a custom script shows how much buyers disliked it. Since .conf21, Splunk Cloud customers can choose workload pricing instead, billed on search compute in Splunk Virtual Compute units. That makes Splunk the one vendor in the set that offers a way out of volume-based billing.

Two meters on one payload. Arize AI’s AX Pro includes 50k trace spans and 10 GB of ingestion per month. Verbose RAG payloads can exhaust the GB ceiling while span count still looks healthy, and high fan-out agents do the reverse. On 2026-07-29 Arize also removed its previously published $3/GB overage rate from the pricing page.

Dropping data instead of billing it. Sentry does not bill usage beyond a customer’s pay-as-you-go budget. It drops the data. That caps the bill and moves the risk from cost to missing telemetry.

What this means for buyers vs vendors

For buyers

• Price the whole path of a gigabyte, not just the ingest rate. Ask what the same GB costs to index, retain for your required window and query. Then model two cases: everything searchable, and a filtered slice.
• Forecast volume from your logging configuration, not your headcount. Size the next 12 months from current GB/day plus planned services, and set ingest alerts well below the allowance. Thresholding and alerting on usage covers the mechanics.
• Ask for the ingest-rate history and a renewal cap in writing. Where the per-GB overage is unpublished or quote-only, get the number contractually before you sign.
• Decide how you want to fail. A hard cap that drops data and uncapped overage are different risks. Pick one deliberately for each data type: security logs usually cannot be dropped, and debug logs usually can.

For vendors

• Pair the ingest meter with a cheap cold path. Customers who can park low-value data at a lower rate log more and churn less than customers who have to choose between visibility and budget.
• Publish the per-GB overage. A self-serve buyer who cannot price the overage treats the plan as enterprise-only. Choosing the right usage metric covers when a volume meter fits and when compute or entity metering is fairer.
• Build the volume-reduction tooling yourself. Customers will filter at the source anyway. Doing it in-product keeps the data relationship and earns trust at renewal. To test how ingest, index and retention rates combine into a bill, use the pricing calculator.

Explore this theme in the knowledge graph →