Datadog pricing for Korea: AP1 rates and a cost worksheet

When a Korean team researches Datadog pricing, a Korean-language page, a Korean hosting region, and a Korean customer’s contract are three different things. This guide builds a cost worksheet for procurement in Korea, using AP1 Japan’s public USD rates as a reproducible regional benchmark. It is not a Korea-hosted rate card or a KRW quote.

This guide is published by Guance, which offers an observability platform discussed below. We checked the vendors’ public rates and billing documentation on 27 October 2026 and recalculated the examples. We did not inspect customer contracts, invoices, or private discounts, or run a production equivalence test. The differences apply only to the disclosed workload; they are not customer quotes or universal savings guarantees.

Datadog pricing: the short answer

Datadog publishes separate annual, monthly, and on-demand rates. On the public AP1 list, Infrastructure Pro starts at US$18.75 per host per month with annual billing, APM at US$38.75 per APM host per month, log ingestion at US$0.125 per GB, and 30-day indexed logs at US$3.125 per million events. These are component rates, not an all-in platform price.

For a useful estimate, collect at least:

• hourly infrastructure and APM host usage;
• containers above included allotments;
• ingested and indexed custom metrics;
• uncompressed log GB and indexed log events by retention tier;
• ingested and indexed spans;
• RUM, Session Replay, synthetic, profiling, network, security, and support usage;
• annual commitment, negotiated discount, overage, tax, and migration cost.

The authoritative starting points are the Datadog AP1 price list, Datadog product allotments, and the account’s actual usage and contract.

A Korean-language price page is not a Korean-region price

Datadog Sites lists AP1 in Japan and AP2 in Australia, but no South Korea site. Sites are independent; data is not shared across them, and some products and ingestion endpoints differ.

The /ko/pricing/list/ path selects a language, not necessarily AP1. A default Korean page can still display a US site, so this guide fixes the benchmark with the official ?site=ap1 URL. This does not require every Korean buyer to use AP1, establish Korean data residency, or make public prices equal to the invoice. Confirm the actual site, currency, discount, and purchase terms in the quote.

Annual, monthly, and on-demand are different buying modes

The AP1 page shows three columns:

Do not mix an annual Infrastructure price with an on-demand log price and describe the result as one standard package. Datadog’s billing documentation and the signed order form should determine how committed, included, and on-demand usage interact. Subscription Details can show contracted quantities and rates, but may not be available for partner-billed accounts; use the order and billing records in that case.

AP1 public rate card at a glance

The following rates were checked against the official AP1 list on 27 October 2026. Currency is USD. Calculations retain the precision of the published unit rates; only displayed subtotals are rounded to cents.

Source: Datadog AP1 pricing. The table is not exhaustive. Database Monitoring, RUM, Session Replay, Synthetics, network products, security products, support, and other capabilities have additional meters.

The APM host rates above use the Infrastructure-attached purchasing model. Datadog’s current APM pricing FAQ also permits standalone APM, APM Pro, and APM Enterprise without Infrastructure Monitoring. Only the attached model requires Pro or Enterprise Infrastructure Monitoring on each APM host. The AP1 FAQ publishes standalone APM at US$45, APM Pro at US$51.25, and APM Enterprise at US$58.75 per host/month, with annual billing and on-demand availability. Verify the purchased mode and quote; the attached rates in the table must not be substituted for standalone prices.

Infrastructure hosts, containers, and high-water usage

Infrastructure Monitoring is not simply “average hosts × headline price.” Datadog documents usage aggregation and allotments in its account billing guide and product allotments page.

Datadog documents hourly host metering and a monthly high-water mark: the maximum usage among the lower 99% of hours. Confirm which Kubernetes nodes, VMs, and App Service Plan instances count as hosts.

For Infrastructure Pro, the public allotment page currently lists, per host:

• 100 Custom Metrics;
• 100 Ingested Custom Metrics;
• 5 Containers;
• 500 Custom Events.

The aggregation period and eligibility differ by meter, and special contract terms take precedence. Containers above the included quantity can produce separate charges. Ephemeral clusters, autoscaling, short-lived build nodes, and installing a separate Agent in each container can therefore change the bill even when average fleet size looks stable. Ordinary Agent and cloud-integration coverage can identify overlapping hosts; two collection paths do not automatically mean double billing. Use Datadog’s billed-usage exports rather than a CMDB monthly average.

Datadog custom metrics pricing needs two counts

Custom-metric costs need their own worksheet because Datadog distinguishes indexed Custom Metrics from Ingested Custom Metrics.

Custom Metrics generally depend on a metric name plus unique tag-value combinations. High-cardinality dimensions such as container IDs, request IDs, customer IDs, or rapidly changing deployment labels can create many billable series. The official Custom Metrics billing guide explains the counting model and included allotments.

The AP1 list publishes:

• US$6.25 per 100 Custom Metrics per month;
• US$0.125 per 100 Ingested Custom Metrics per month.

Datadog also documents a separate Metric Name Pricing model. Its applicable commercial rate is not published on the reviewed page, so it remains unconfirmed for the account until its contract and usage method are verified. An unconfirmed commercial input does not mean the capability is unavailable or the technical explanation is defective.

For the cardinality-based model, Ingested Custom Metrics charges apply to metrics whose tags or aggregations are configured with Metrics without Limits™. Unconfigured metrics are charged for indexed volume only; do not add both meters indiscriminately. Export the applicable counts before estimating. If the account uses Metric Name Pricing instead, use that contract’s metric-name and data-point model rather than these series-based rates.

APM pricing combines hosts, ingestion, and indexed spans

The AP1 annual list price for base APM is US$38.75 per APM host per month. Datadog’s APM billing documentation and product allotments currently show an included allowance of 150 GB of ingested spans and one million indexed spans per APM host per month for base APM.

Above the applicable allowance, the AP1 list publishes:

• US$0.125 per ingested GB;
• US$1.5875, US$2.125, or US$3.125 per million indexed spans for 7-, 15-, or 30-day retention under annual billing.

The cost can change with Kubernetes node counting, how the Agent is deployed, service sampling, span volume, retention, and the number of spans deliberately kept for search and analytics. “We send OpenTelemetry” does not remove these backend meters.

Log pricing has both ingestion and indexing

Datadog log cost commonly contains at least two separate components:

1. ingestion for the uncompressed data sent to Datadog; and
2. indexing for the events retained in an index for the selected period.

The AP1 annual public rates used in the model below are:

• US$0.125 per GB of ingested logs;
• US$3.125 per million indexed log events at 30-day retention.

Datadog’s Log Indexes documentation explains that exclusion filters can control which logs remain indexed. Excluding an event from an index can reduce indexing usage, but it does not retroactively remove the ingestion that already occurred. Archives, rehydration, Flex Logs, forwarding, sensitive-data scanning, and pipelines can add other meters.

This distinction is a major reason a “price per GB” comparison can be incomplete: one vendor may charge for bytes plus retained events, while another may use a mutually exclusive entry or traffic model.

RUM, Session Replay, Synthetics, and support must be added separately

An end-to-end Datadog estimate should also identify:

• Browser and Mobile RUM sessions;
• Session Replay sessions and retention;
• API and browser synthetic tests;
• profiling, network, database, security, and incident-management products;
• the selected support plan.

The AP1 list contains the applicable product meters, while Datadog support plans describe Free, Standard, and Premier support structures. Support percentage, response targets, and eligibility can change the commercial result and must come from the quote.

Build the Datadog cost worksheet before comparing vendors

Use one row per product meter and keep committed, included, and on-demand quantities separate.

Keep public and contract rates in separate columns, and label each input as an official source, completed experiment, signed contract, or unconfirmed assumption. No customer contract or completed experiment is supplied by this article. The comparison should then apply each platform’s units independently. Equal source data does not guarantee equal billed usage.

A reproducible 30-day log scenario

This narrow scenario is intentionally limited to logs because the public meters can be aligned more transparently than a mixed Infrastructure, APM, RUM, and Synthetic bundle.

Assumptions:

• 1 decimal GB of original, uncompressed logs per day;
• 1 million log events per day;
• every event is indexed and retained for 30 days in Datadog;
• every event is under both Guance large-entry splitting thresholds;
• a steady 30-day billing month;
• Datadog AP1 annual public rates;
• Guance Hong Kong & Global default data-count rate;
• no discount, support, tax, archive, rehydration, forwarding, pipeline, scanning, migration, or feature-equivalence claim.

Guance’s official price is US$0.60 per million billed log entries per day at 30-day retention. The Guance billing logic says data-entry billing is the default log model. It also documents that events and self-built synthetic data can enter the same billing item, custom indexes are counted separately, and large entries can be split. The documentation gives large-entry thresholds of 10 KB for ES and 2 KB for SLS; verify the applicable backend. Those additions are excluded from this controlled example.

In this disclosed log-only public-rate scenario, Guance is 81.5% lower than the Datadog AP1 annual-list result. That is not a universal Datadog-to-Guance saving and does not establish equal features, search performance, support, or contract scope.

Formula and calculation method

For this scenario:

Datadog monthly logs
= (daily uncompressed GB × 30 × US$0.125)
+ (daily indexed events in millions × 30 × US$3.125)

Guance monthly logs
= daily billed entries in millions × 30 × US$0.60

Scenario-specific saving %
= (Datadog modeled cost - Guance modeled cost)
  / Datadog modeled cost × 100

At 1 GB and 1 million events per day:

Datadog = (1 × 30 × 0.125) + (1 × 30 × 3.125)
        = US$97.50

Guance  = 1 × 30 × 0.60
        = US$18.00

Difference = (97.50 - 18.00) / 97.50
           = 81.5%

This model assumes an average of one decimal kilobyte per event and, more importantly, that no individual Guance event crosses the documented splitting threshold. Use actual event-size distribution in a customer calculation.

Indexing-ratio sensitivity and the 50% threshold

Scaling the same 100% indexing assumption across traffic volumes would keep the percentage constant. The more useful sensitivity variable is the share of one million daily events kept in Datadog Standard Indexed Logs for 30 days. This table fixes source volume at 1 GB and one million events per day; Guance bills all one million entries under the default data-count model.

Percentages use the unrounded costs. This is not a feature-equivalent comparison: at 10% indexing, Datadog retains only one tenth of the events in the modelled Standard index while the Guance subtotal still counts all entries. The reversal is nevertheless important. It proves that indexing ratio, Standard versus Flex Logs, event-size distribution, Guance entry splitting, and required search scope must be explicit calculator inputs. Discounts, commitments, archives, rehydration, add-ons, and contract terms can move every result again.

What about Guance traffic billing?

Guance also publishes a Hong Kong & Global 30-day rate of US$0.17 per GB of original log write traffic per day. Its official Commercial Plan pricing details and billing logic state that traffic billing is not the default and requires an account manager to enable it.

We exclude that conditional mode from the worked model and headline. Eligibility, contract activation, the definition of GB, entry-size distribution, and the selected commercial terms remain unconfirmed. A buyer can add it only after the account manager confirms the mode in the proposed order.

Guance’s data-count and traffic models are alternatives for standard log storage; they are not added together in this example.

Can Guance be about 50% cheaper than Datadog?

The public evidence does not establish an approximately 50% difference for a mixed observability workload, so the blanket statement “Guance is 50% cheaper than Datadog” should not be used. In this narrow log model, the arithmetic crosses 50.0% when 34.4% of one million daily events are kept in the Datadog Standard 30-day index. It reaches 81.5% at 100% indexing and reverses when only 10% is indexed. None of those points establishes full-platform total cost or feature parity.

The defensible claim is:

As of 27 October 2026, under the disclosed AP1 annual public-rate log scenario with 1 GB and one million events per day, 100% Standard indexing, and 30-day retention, the modeled Guance known subtotal is US$18.00 versus US$97.50 for Datadog, or 81.5% lower. This result applies only to this scenario and is not a customer quote, full TCO, feature-parity claim, or universal saving.

Use the customer’s billed usage and current quotes to determine the direction and magnitude of the full scoped result without presupposing a 50% outcome.

Keep the workload, rate date, site, retention, indexing ratio, billing mode, and exclusions attached to any percentage used outside this article.

When Datadog may cost less or fit better

Datadog may cost less or fit better when:

• a negotiated commitment or volume discount materially reduces the public AP1 rates;
• the organization already has unused Datadog committed spend;
• exclusion filters keep only a small fraction of ingested logs indexed;
• Datadog-specific workflows, integrations, security, support, or team expertise replace other tools or operating labour;
• the Guance account cannot use the intended billing mode;
• Guance billed entries increase because of large-log splitting, multiple indexes, events, or self-built synthetic data;
• Guance add-ons such as forwarding, sensitive-data scanning, security scanning, or Central Pipeline materially increase cost;
• migration, retraining, historical access, and dual-running cost exceed the rate-card difference;
• AP1 or AP2 operational coverage is a better fit than the proposed Guance architecture.

Conversely, Guance may have a stronger cost case when the workload maps cleanly to its published billing unit, the necessary product workflows pass a controlled evaluation, and the applicable commercial and data terms are confirmed.

Procurement in Korea: region, Korean support, and data locations

AP1 is in Japan. Datadog’s Japan data-center announcement describes Japanese storage and processing; it does not establish hosting in Korea or compliance with a particular Korean regulation.

The Korean support documentation lists Korean-language support on weekdays, 09:00–17:00 KST, excluding local holidays, and Korean chat on business days from 10:00–11:30 and 13:00–16:00 KST. Confirm response targets, escalation, and Standard/Premier scope in the order. KRW billing, exchange-rate date and method, VAT, reseller and regional charges, minimum and ramp commitments, support, and onboarding are also quote-specific. We do not make an arbitrary USD-to-KRW conversion.

Guance’s commercial sites include AP1 on AWS Singapore, AP2 on Alibaba Cloud Singapore, ID1 on Tencent Cloud Jakarta, and CN6 on Alibaba Cloud Hong Kong. No Korea or Seoul SaaS site is listed. Accounts and data are independent across sites and cannot be shared or transferred between them. Site names alone do not establish regulatory fit, latency for Korean users, backup/DR/support-data locations, or feature parity.

Get written confirmation from both vendors of the tenant and ingestion endpoints; all relevant data locations; cross-border support access and subprocessors; regional products, currency, tax, reseller terms, SLA, support language and KST coverage, escalation, and implementation ownership; and export, archive, deletion, and exit terms. Guance’s ticket service publishes 5×8 service, not proof of a Korean-speaking team or contractual response target. If deployment within Korea is mandatory, assess the Deployment Plan separately rather than applying SaaS rates and operating assumptions.

Measure both platforms before making a savings claim

OpenTelemetry can support a controlled dual-write for standard signals. Datadog publishes a path-specific OpenTelemetry compatibility matrix, Guance documents OpenTelemetry ingestion through DataKit, and the upstream OpenTelemetry Collector can export to more than one backend.

Run a 7–14 day production-shaped measurement:

1. Export 30 days of Datadog billed usage and the contracted rate schedule.
2. Choose representative services with normal and peak telemetry patterns.
3. Dual-write only the signals whose semantics and privacy controls are understood.
4. Record accepted, rejected, sampled, dropped, ingested, indexed, retained, and queried data.
5. Rebuild only the dashboards and alerts needed to test real incident tasks.
6. Compare investigation coverage, latency, access control, support, and operating labour.
7. Recalculate both invoices with the observed meters and binding quotes.
8. Keep the current alert path and rollback until acceptance criteria are met.

OpenTelemetry compatibility does not migrate Datadog dashboards, monitors, SLOs, roles, workflows, history, or proprietary behavior automatically.

For product-selection context, use the Korean Datadog alternatives guide. For technical acceptance, first use the Korea DDTrace-to-Guance canary Quickstart on one non-production service to check the receiver, fields, sampling, investigation workflow, and rollback. It does not prove migration of dashboards, monitors, SLOs, logs, RUM, or history.

The proposed experiment has not been performed for this article. A documented procedure is not evidence of a successful test or approval for a production cutover.

Procurement checklist

Before accepting a Datadog quote, ask:

1. Which site and rate card apply to the proposed account?
2. What are the annual commitment, monthly, on-demand, and overage rates for every used product?
3. Which host, container, custom-metric, log, and span allotments apply?
4. How are high-water, hourly, and monthly quantities calculated for this contract?
5. What were the measured ingested GB and indexed events or spans during the POC?
6. Which retention tiers and archives are included?
7. What support, tax, reseller, implementation, and regional charges are excluded?
8. What happens to unused commitment and usage above commitment?
9. Which products or features differ between AP1 and AP2?
10. What export, rehydration, and termination charges apply?

Before accepting a Guance quote, ask:

1. Does the Hong Kong & Global public rate apply to the proposed site in the order form?
2. Is log billing based on default billed entries or enabled ingest traffic?
3. How are large events, multiple indexes, events, and synthetic data counted?
4. Does “GB” mean decimal GB or GiB in the contract?
5. Which forwarding, scanning, pipeline, trigger, replay, network, and support charges apply?
6. Which Datadog assets require manual recreation?
7. What billable quantities did the dual-write POC produce?
8. Which data, backup, support, DPA, SLA, and exit terms are binding?

Unconfirmed inputs, exclusions, and updates

This model excludes negotiated discounts, annual-commitment mechanics, unused spend, support, tax, reseller terms, archives, rehydration, forwarding, pipelines, scans, RUM, synthetics, network products, security products, APM, Infrastructure, migration labour, training, data egress, and feature or performance equivalence.

The following remain unconfirmed until measured or quoted:

• AP2 public and contracted rates;
• the actual Datadog site proposed for the buyer;
• the binding rate schedule for the proposed Guance site;
• Guance GB unit definition and billing-mode approval;
• the customer’s actual included allotments and discounts;
• how both vendors count the same real telemetry after filtering, sampling, splitting, and indexing;
• whether the required workflows, support, SLA, and regional terms are equivalent enough for the customer.

Pricing verified 27 October 2026. Recalculate when rates, allotments, aggregation, retention, entry-splitting rules, regional availability, contract conditions, or measured quantities change. The excluded items define the scope of the model; they are not evidence of product defects.

Official sources

• Datadog: AP1 detailed pricing, default detailed pricing, Sites, billing, product allotments, container billing, APM billing, custom metrics, Metric Name Pricing, log indexes, OpenTelemetry compatibility, and support plans.
• Guance: Commercial Plan pricing details, billing logic, commercial sites, and OpenTelemetry integration.
• OpenTelemetry: Collector documentation.