Datadog pricing in Japan: hosts, APM and log costs

For a Japanese Datadog deployment, I would decide which data needs to remain searchable before multiplying a host price. Logs make the distinction especially clear: ingested bytes, indexed events and retention each affect the bill. Keeping every event searchable for 30 days is a different purchase from indexing only the most useful 10%.

Datadog AP1 is a strong starting point when a Japanese SaaS location is a requirement. Teams already using its application views and alerts also have good reasons to improve indexing and metric cardinality before moving. Guance becomes relevant when retaining a broader set of logs matters and indexing costs constrain that scope. Its published SaaS locations do not include Japan, however, so a domestic-deployment requirement must be settled before comparing the rates.

Guance Editorial Team. Official prices and documentation checked on 27 October 2026. Guance supplies one of the products discussed here. The figures are public-rate calculations for stated assumptions, not customer invoices or performance measurements.

AP1 Japan and the price list that applies

Datadog Sites lists AP1 in Japan, with the console at ap1.datadoghq.com. Sites are independent. Check the account's actual site rather than the language of the page; this article uses the AP1 rate card explicitly selected with ?site=ap1.

A Japanese location is a concrete advantage for teams that want telemetry stored and processed domestically. Backup, support access and product-specific coverage remain contractual matters. A yen budget should likewise use the actual billing currency, conversion rules and consumption tax rather than an arbitrary exchange rate.

Host rates and the two ways to buy APM

These selected list rates are in USD. Annual-billing prices are still expressed per month, not as an annual total. The invoice applies the purchased commitment, included allowances and overage terms.

The US$38.75 APM host rate is for the Infrastructure-attached model, which requires Pro or Enterprise Infrastructure Monitoring on that APM host. The AP1 pricing FAQ also lists standalone APM without Infrastructure: US$45 for APM, US$51.25 for APM Pro and US$58.75 for APM Enterprise per host/month.

At annual list rates, one host with Infrastructure Pro and base APM therefore starts with 18.75 + 38.75 = US$57.50/month for those products. Standalone APM is US$45/month. The former also buys infrastructure monitoring, so the choice depends on that requirement. Neither sum includes logs, data overage or other products.

How hosts and metric series grow matters

Infrastructure host billing is not simply the monthly average. Datadog measures usage hourly and uses the highest value in the lower 99% of hours as its monthly high-water mark. A Kubernetes node that exists only during the working day should not be assumed to contribute a negligible amount. Host billing

The standard Infrastructure Pro allotments list 100 Custom Metrics, 100 Ingested Custom Metrics, five containers and 500 Custom Events per host. Base APM includes 150 GB of ingested spans and one million indexed spans with default 15-day retention per host/month. Apply the relevant allowance before pricing excess quantities.

Under cardinality-based Custom Metrics pricing, different tag-value combinations create distinct series. Customer IDs and request IDs can turn a few metric names into a large population. Ingested Custom Metrics metering applies to metrics configured with tags or aggregations through Metrics without Limits™; unconfigured metrics should not automatically be charged under both ingestion and indexing. Custom Metrics billing

The separate Metric Name Pricing contract model uses names and data points, so the series-based rates do not apply. APM also distinguishes spans received from spans retained for search. Above applicable allowances, annual indexed-span rates per million are US$1.5875 for seven days, US$2.125 for 15 days and US$3.125 for 30 days. APM billing

Logs are billed for ingestion and indexing separately

Datadog log ingestion counts uncompressed GB; Standard Indexed Logs counts events and their retention period. AP1 annual list rates are US$0.125/GB ingested and US$3.125 per million events indexed for 30 days. Log index configuration

Exclusion filters reduce the share of received logs kept in an index. Ingestion has already happened, so reducing that charge requires filtering before transmission. Frequently investigated records can remain indexed while less-used data follows another path. Flex Logs, archives and rehydration have their own meters, which belong in any estimate using those routes.

Guance's standard-log model offers a different calculation. Its default data-count rate for Hong Kong & Global is US$0.60 per million billed entries added per day at 30-day retention. This makes the amount of history a team wants to keep a useful starting point for comparison, especially when a small indexed subset would not cover its investigations.

Billed Guance entries need not equal application log lines. Large-entry splitting thresholds are 10 KB for ES and 2 KB for SLS; multiple indexes, events and self-built synthetic data can also affect quantities. The following example assumes none of those adds entries. Billing logic

One million daily events, retained for 30 days

Assume one decimal GB and one million events each day for 30 days, with every entry below Guance's splitting thresholds. Datadog uses AP1 annual list prices with all events in a Standard index. Guance uses the default Hong Kong & Global data-count model. Tax, support, discounts, archives, rehydration, forwarding, pipelines, scanning and migration are excluded.

Of Datadog's US$97.50 subtotal, US$93.75 comes from indexing. Choosing which records need immediate search has more influence here than shaving a little off their byte size. Guance's US$18.00 subtotal is 81.5% lower under these assumptions. That is not a whole-platform comparison covering APM, RUM, Infrastructure, query performance or operations.

Let G be daily uncompressed GB, E daily events in millions, and p the fraction indexed in Datadog. Guance retains all entries in this example, without additional splitting or other entry charges:

Datadog = 30 × (G × 0.125 + E × p × 3.125)
Guance  = 30 × E × 0.60

Change the indexing ratio and the result reverses

Hold the source at one GB and one million daily events, varying only Datadog's indexing fraction. Guance continues to count all entries.

The equal-cost point is 15.2% indexing: 3.75 + 93.75 × 0.152 = 18. The 34.4% row is a different threshold: Datadog reaches US$36 and Guance is half that amount. At 10%, Datadog costs less, but its Standard index also holds only one tenth of the events.

Decide the required search coverage first. A team needing every event retained and one needing immediate search over a small subset can reach different conclusions from the same rate card. Contract discounts, Flex Logs and Guance's actual post-splitting counts move the crossover again.

Guance also lists a 30-day traffic rate of US$0.17 per daily GB written. It requires account-manager activation and is not the default. Traffic and entry billing are alternatives; use the mode confirmed in the order rather than adding them together.

When keeping Datadog is worth the cost

A Japanese SaaS location, working alerts and investigation views, and an experienced team are substantial reasons to keep AP1. If committed spend remains unused, adding another service may not reduce cash spending immediately even at a lower unit price. Improving indexing, retention and cardinality is often the more direct saving.

Guance is worth comparing when broad log retention and hosted application correlation are the main requirement. But its published AP1 location is AWS Singapore, not Datadog's AP1 Japan, and its SaaS list does not include Japan. If domestic deployment is mandatory, assess the Deployment Plan separately with its infrastructure and operating costs; the SaaS rates above cannot price that arrangement.

Saved queries, dashboards, alerts and historical access are also assets. The Japan DDTrace integration guide covers an existing instrumentation path, but protocol support does not automatically move those assets. If rebuilding them consumes the rate difference, a lower list price alone is a weak reason to replace the platform.

Finally, add the RUM sessions, replays, API and browser tests, network and security products, and support service actually used. Public rates explain how the cost forms. The purchase should follow the data the team needs to retain and the investigations it performs every day.