When platform and FinOps leaders ask what Datadog costs for CDN analytics, they are usually trying to forecast a line item that does not exist on the price card. CDN visibility is a use case, and Datadog sells modular observability products that can support pieces of that use case, each with its own billing unit.
That mismatch is why quotes vary so widely across accounts with similar traffic. Two teams can both monitor the CDN in Datadog and land on very different monthly spend, depending on whether full request logs go to Standard Indexing, Flex Logs, or archive for later scan, or whether the stack only pulls aggregated metrics.
Key Takeaways
- Datadog does not publish a CDN analytics product line. The use case is assembled from log management, metrics and integrations, and optional add-ons.
- The expensive meters at edge volume are log ingest per GB, Standard indexed events times retention, and custom-metric cardinality. Flex Logs Storage ($0.05 per million events stored per month) and Archive Search (in-place archive queries) are the current cost-control path, not a free pass.
- Full request logs at hundreds of thousands to millions of requests per second make list-price models from app hosts and APM fail, so which tier holds the edge stream (Standard, Flex, or archive-scan) is the multiplier.
- Before a forecast, specify metrics-only versus hot indexed logs versus Flex versus Archive Search versus a dedicated store, because those choices change which meters fire and how interactive the investigation is.
- Teams that need months of unsampled, always-queryable edge logs often keep Datadog for application operations and put CDN request logs in a dedicated analytics layer so observability list prices and scan-based archives do not set investigation policy.
There Is No Single CDN Analytics SKU
Datadog’s public catalog is organized around infrastructure monitoring, APM, log management, RUM, synthetics, and cloud cost management, and none of those is labeled CDN analytics.
What operators mean in practice usually falls into three buckets: edge metrics and health; request-level logs; and CDN spend analytics, where Cloud Cost Management attributes provider invoices. Only request logs behave like forensics-grade CDN analytics, because metrics show that something changed, while logs show which clients, paths, status codes, and cache statuses were involved.
If an RFP says “CDN analytics,” force the vendor conversation onto which of those three is in scope. Native reports that stop at summaries are why teams that need request-level evidence export.
The Meters That Drive Cost
For CDN-scale work, the expensive meters sit in how logs are collected, stored, and queried. Confirm current list and contracted rates, because published numbers are the shape of the cost function, not a quote.
Log Ingest at CDN Volume
Datadog charges for logs ingested. Public list is $0.10 per uncompressed GB ingested per month (annual, month-to-month, and on-demand), and high-traffic CDNs can produce hundreds of gigabytes to multiple terabytes of raw logs per day. Ingest applies whether the stream is indexed, sent to Flex, or archived, so filtering and sampling before ingest are the only ways to shrink this meter.
Indexed Events and Retention
Searchable logs in Standard Indexing are billed separately, per million events, with retention as the price ladder. On the public annual list, that is $1.06 per million events at 3-day retention, $1.27 at 7-day, $1.70 at 15-day, and $2.50 at 30-day. Month-to-month and on-demand are higher (30-day is $3.00 and $3.75), which is why moving from 15 days to 30 is often where this line item becomes a board-level number.
Flex Logs and Archive Search
Datadog’s 2026 cost story for CDN logs is not “index everything or rehydrate cold archives.” Flex Logs has been generally available since 2024 (public posts dated 20 September 2024). The public list splits Flex Logs Storage at $0.05 per million events stored per month (annual) from Flex Logs Starter at $0.60 per million events for 6-, 12-, or 15-month packages (annual), and Flex docs also mention a 3-month Starter option. Datadog’s Flex documentation lists CDN services (Akamai, Cloudflare, Fastly, and CloudFront) as recommended direct-to-Flex sources.
On 26 June 2026 Datadog published a CDN cost playbook on reducing CDN log costs with searchable archives. Observability Pipelines can route raw edge logs to customer object storage in Datadog’s archive format, and Archive Search, generally available since 10 June 2025, runs queries on those archives in place without rehydrating the store first. An optional Search and Rehydration mode can still index a result set into Log Explorer. Scan compute, possible cloud egress, and a preview cap of about 100,000 results retained for 24 hours remain on the bill.
Flex and Archive Search price infrequent or after-the-fact edge investigations, and they do not make months of unsampled CDN logs as cheap or as interactive as a store built for that workload. The RFP should document the split (Standard Indexing, Flex, Archive Search) and ask whether investigations must be sub-second on petabytes or whether a scan-and-preview is acceptable.
Custom Metrics and Cardinality
CDN integrations and custom metrics can look cheap next to logs until cardinality grows. Public list is $5 per 100 custom metrics per month (annual), plus $0.10 per 100 ingested custom metrics, and hostname, status class, cache status, and PoP tags on an edge stream produce a cardinality explosion that is easy to miss in a pilot. Those series can dominate the bill if high-dimension metrics are emitted to avoid log volume.
Add-Ons Left Out of the Model
Cloud Cost Management can help attribute CDN provider spend. Public list for Enterprise is $10 per $1,000 in cloud or SaaS spend per month (annual), which is FinOps work, not request-log retention, and it adds a meter. Forwarding to custom destinations is listed at $0.25 per GB outbound per destination per month, and Infrastructure Pro starts at $15 per host per month billed annually.
Why CDN Volume Outruns App-Log Pricing
Teams that model Datadog from hosts, APM, and a few hundred GB of app logs, then add CDN Logpush, often discover that the edge stream is larger than everything else combined.
A directional list-price sketch (not a quote): 1 TB per day of uncompressed CDN logs is about 1,024 GB per day times $0.10, or roughly $3,070 per month of ingest before indexing, Flex, or archive-scan. If the average record is about 1 KB, that is on the order of a billion events per day, so indexing even a fraction at 15-day Standard rates dominates ingest. Confirm contracted rates.
Cost pressure then shows up as sampling, exclusion filters, or moving the bulk of the stream off Standard Indexing. Those controls reduce the invoice while changing what can be proved during an incident. Launch-scale edge traffic is the extreme case, because traffic can spike 25x, multi-CDN fabrics can emit millions of log lines per second, and a 2 percent regional error rate is easy to sample away.
If the requirement is months of full-fidelity edge data without a scan wait, pricing the problem only as another Datadog log source will push the design toward Flex, archive-scan, or sampling, which is why separating status-page metrics from request-log questions comes before budget.
What to Verify Before Committing Budget
Ask for a written cost model that includes peak and average requests per second, average bytes per log record after transform, and whether logs are sampled, filtered before ingest, or fully retained. State which CDN streams go to Standard Indexing, Flex Logs, and archive, and include Archive Search scan compute and cloud egress.
Specify CDN log retention, not the account default, and state whether investigations must be sub-second on petabytes or can wait on a scan-and-preview. Multi-CDN scope usually means more volume, which is the problem correlating providers in one view is designed for, and the model should state who owns the bill when a spike multiplies ingest for a week.
Separate “the CDN is visible in Datadog today” from “last quarter’s edge questions can be answered without a scan window.” What to measure at the edge and using the same stream for security and performance help clarify which questions need raw logs versus dashboards.
Map hot, warm, and cold storage onto Standard, Flex, and archive in the RFP so procurement is not comparing a metrics tile to a forensics store.
When a Dedicated Edge Data Layer Changes the Math
Some organizations keep Datadog for application observability and route CDN request logs into a purpose-built store for high-volume, long-retention analytics, so observability meters and archive-scan cost do not set policy for edge forensics.
That pattern shows up when the requirement is a unified view across providers, months of full-fidelity CDN logs, and query performance that does not require thinning the dataset first. Aggregates can still export to Datadog for dashboards and alerts on-call already trusts.
Hydrolix is built for that split: keep Datadog as the system of engagement, and put the edge stream on usage-based billing designed for terabyte-scale retention without sampling. Cost is the sum of the meters that are turned on, dominated by ingest plus whichever of Standard, Flex, or Archive Search is assigned to the request-log stream.
Walk Through CDN Analytics Cost with Edge Traffic
If a team is forecasting CDN analytics spend under real edge volumes and retention targets, a walkthrough can show how operators separate operational tooling from long-retention request analytics. Book a demo to pressure-test the model with stated traffic assumptions.
Does Datadog sell a CDN analytics product?
No. CDN analytics is a use case assembled from logs, metrics, integrations, and sometimes RUM or synthetics, so quotes only compare if the RFP states which of those is in scope. Native hourly summaries and a full request-log pipeline are different products in practice.
Is Cloud Cost Management the same as CDN analytics?
No. Cloud Cost Management helps attribute what an organization pays CDN and cloud providers. It does not replace retaining and querying full request logs for performance or security investigations.
What usually dominates the Datadog bill for CDN logs?
Ingest volume plus whichever of Standard Indexing, Flex Logs, or Archive Search is assigned to the edge stream. At CDN scale, request logs often exceed application log volume, so the tier and retention become the main budget lever. Confirm current list and contracted rates.
How do Flex Logs and Archive Search change cost?
Flex Logs (generally available since 2024) prices in-Datadog retention as Flex Storage at $0.05 per million events stored per month, or Flex Starter at $0.60 per million events for 6-, 12-, or 15-month packages, with a 3-month Starter option also mentioned in Flex docs. Datadog recommends CDN sources such as Akamai, Cloudflare, Fastly, and CloudFront for direct-to-Flex. Archive Search, generally available since 10 June 2025, queries object-storage archives in place without rehydrating first, and neither path is the same as always-hot, sub-second query on months of unsampled edge data.
How should teams compare Datadog to a dedicated store?
Compare the same requirements: unsampled ingest rate, retention horizon, multi-CDN schema normalization, and query latency on raw data. Do not compare a metrics dashboard or an archive scan to a full-fidelity log store.
