Datadog Pricing in Singapore: Logs, Custom Metrics, and AP1 Costs

If a Singapore team’s Datadog bill is growing, I would first separate log ingestion, indexed events and custom metrics. They respond to different changes. Keeping fewer logs in the Standard index can lower one charge without reducing ingestion; adding a customer-ID tag can increase metric series while host count stays flat. A lower headline price will not explain either movement.

This Datadog pricing guide uses AP1 Japan’s published USD rates to make those calculations visible. It is an APAC reference for a Singapore buyer, not a Singapore-hosted service price or an AP2 Australia quote. The account’s site and contract still determine the rates it pays.

Guance Editorial Team · 27 October 2026. Guance publishes this article and appears in the log-cost example. The figures are calculated from official rate cards, not customer invoices or production tests.

AP1 rates and the APM purchase model

The table uses Datadog’s AP1 Japan public list, checked on 27 October 2026, in USD. Annual rates depend on the relevant commitment; monthly and on-demand columns describe different purchasing terms. Included quantities, negotiated rates, unused commitments and overage belong in the account’s calculation. Keep the published precision until rounding the final subtotal.

The APM rates in this table are for purchasing APM with Infrastructure Monitoring. That model requires Pro or Enterprise Infrastructure on the APM hosts. Standalone APM is also available: the AP1 FAQ lists 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. Comparing the attached APM rate with a standalone quote leaves out a material part of the purchase.

Hosts, spans and custom metrics grow differently

Infrastructure charges need the account’s hourly billed usage rather than an average VM count. Scaling nodes and excess containers can change the bill. The standard Pro allotments include 100 Custom Metrics, 100 Ingested Custom Metrics, five containers and 500 Custom Events per host, subject to each meter’s aggregation rules and the contract. Normal Agent and cloud-integration host recognition can deduplicate overlapping hosts; two collection paths do not by themselves prove two charges.

APM then adds its own volume. Base APM includes 150 GB of ingested spans and one million indexed spans per host/month with 15-day default retention under the published allowances. Extra ingestion uses the AP1 US$0.125/GB rate; annual indexed-span rates are US$1.5875, US$2.125 and US$3.125 per million for 7, 15 and 30 days respectively. Sampling and retention change these quantities even with a stable host count. APM billing

For Datadog custom metrics pricing, count the series created by metric names and tag-value combinations. A new customer-ID or container-ID dimension can increase billable series without adding a metric name. AP1 lists US$6.25 per 100 Custom Metrics/month and US$0.125 per 100 Ingested Custom Metrics/month, above applicable allowances. The ingestion charge applies to metrics configured with Metrics without Limits™; unconfigured metrics are charged for indexed volume only. A Metric Name Pricing contract instead uses metric names and data points, so the series-based calculation does not apply. Custom Metrics billing

RUM sessions, replay sessions, API tests and browser tests also have distinct meters. Database, network, security, profiling and support products add their own costs. Keeping these quantities separate explains a growing bill more reliably than multiplying every item by traffic growth.

Log cost follows the events you keep searchable

Datadog charges for uncompressed log ingestion and for events retained in the selected index. An exclusion filter reduces indexed events; it does not undo ingestion already performed. That makes the indexing ratio a useful cost lever when only a fraction of records needs Standard indexed search. Flex Logs, archives and rehydration provide other arrangements with their own charges. Log indexes

Guance’s Commercial Plan uses a different unit. Its Hong Kong & Global public rate for 30-day retention is US$0.60 per million new billed log entries per day under default data-count billing. Billing logic includes large-entry splitting: the documented thresholds are 10 KB for ES storage and 2 KB for SLS. Custom indexes are counted separately; events and self-built synthetic data can also enter this billing item. The source event count is therefore only a starting point.

Consider a deliberately narrow, steady 30-day example: 1 decimal GB and one million events each day, all indexed in Datadog Standard for 30 days, with every event below both Guance splitting thresholds. Use AP1 annual list rates and Guance’s Hong Kong & Global default data-count rate. Exclude discounts, commitments, support, tax, reseller and currency effects, other products, archives, rehydration, forwarding, pipelines, scans, export, migration and parallel running. This is a rate calculation, not a production test or a feature-equivalence result.

The inputs can be changed independently:

D = daily uncompressed GB × 30 × 0.125
  + daily indexed events in millions × 30 × 3.125
G = daily billed entries in millions × 30 × 0.60
Difference (%) = (D - G) / D × 100

D = 3.75 + 93.75 = US$97.50
G = US$18.00
Difference ≈ 81.5%

The 81.5% difference belongs to this log-only scenario. An average event size of 1 decimal KB does not guarantee that every event is below a splitting threshold; the individual size distribution still matters.

A lower indexing ratio can reverse the result

Keep the same source volume and let Guance count all one million daily entries. Only change the share retained in Datadog’s Standard index:

Percentages use the unrounded amounts. At 34.4% indexing, Datadog’s US$36 subtotal is twice Guance’s US$18. At 10%, Datadog is lower. The latter retains only a tenth of the source events in the modeled Standard index, so the searchable scope differs. A useful decision starts with the records the team needs to retrieve, then prices that scope.

Guance also publishes US$0.17 per GB of daily original write traffic at 30-day retention for Hong Kong & Global. This alternative requires account-manager activation. It replaces default entry billing for standard log storage rather than being added to it. Its availability, GB definition and applicable site rate need to be confirmed in the order before using it in a budget.

AP1 means Japan at Datadog and Singapore at Guance

The same region code can hide a substantial architectural difference. Datadog’s site list places AP1 in Japan and AP2 in Australia, with no Singapore site listed. Guance’s commercial sites place AP1 on AWS Singapore and AP2 on Alibaba Cloud Singapore. A quote that says only “AP1” has not yet told you enough.

For a Singapore workload, a Guance Singapore site is a concrete local hosting option to consider alongside Datadog’s proposed site. It does not establish latency, complete data residency or regulatory compliance by itself. Production storage, backups, disaster recovery and support-data access need their own agreed locations and terms. Accounts and data are independent between Guance sites; its documentation does not provide shared or transferable accounts across them.

The price also needs to follow that choice. Confirm whether Guance’s Hong Kong & Global public schedule applies to the Singapore order. Datadog AP1 rates must not be reused for AP2 without the appropriate quote. USD or SGD billing, tax, reseller arrangements, support, commitment, unused spend and overage can change the payable amount without changing any telemetry.

Reduce avoidable volume before paying for a migration

For a team already working effectively in Datadog, I would start with indexed-log scope and high-cardinality tags. Those changes can reduce recurring usage while keeping dashboards, alerts and engineering habits intact. A negotiated rate or unused commitment may further reduce the value of moving now.

Guance is worth comparing when the team wants a managed platform available in Singapore and its log volume fits the proposed entry or traffic model. Keep the required investigation history in that comparison. A cheaper subtotal that leaves out records engineers need to search has not bought the same service.

Migration adds dashboard and alert reconstruction, historical access, training and a period of parallel operation. Standard OTel ingestion does not move those assets automatically. Use the actual billed quantities and the two proposed service scopes to decide whether the recurring difference covers that work. For the broader product decision, see Datadog alternatives for Singapore.