Datadog vs Dynatrace: Per-Host Pricing, Grail, and Which Fits Your Stack
Prices and facts checked on vendor pages on Nov 8, 2026.
Datadog and Dynatrace both cover infrastructure, APM, logs, real user monitoring, synthetics, and security in one product. Datadog sells each capability as its own line item and leans on a large integration catalog. Dynatrace sells a single annual platform commitment, installs one agent per host, and stores everything in its Grail data lakehouse.
The Short Answer
Pick Datadog if you want to buy product by product, start on a free tier or month-to-month plan, and mix its Agent with OpenTelemetry as you like. Pick Dynatrace if you run larger, long-lived estates where automatic discovery and topology-driven root cause analysis matter, and you are comfortable committing spend at the platform level. The biggest budget difference is how hosts are metered: Datadog per host, Dynatrace per GiB of host memory.
Datadog vs Dynatrace at a Glance
How Datadog Charges
Datadog's list prices page shows three columns: annual, monthly, and on-demand. Annual rates checked Nov 8, 2026:
- Infrastructure Pro is $15 per host per month ($18 monthly or on-demand), and Infrastructure Enterprise is $23.
- APM is $31 per APM host per month. APM Pro is $35, and APM Enterprise is $40 and includes Continuous Profiler.
- Each APM host includes 150 GB of ingested spans and 1 million indexed spans a month, summed across all your APM hosts. Extra ingestion is $0.10 per GB.
- Logs cost $0.10 per GB ingested. Indexed events are $1.06, $1.27, $1.70, or $2.50 per million for 3, 7, 15, or 30 days of retention.
- Flex Logs Storage is $0.05 per million events stored. Flex compute is a separate tier, and monitors and Watchdog do not run on Flex Logs.
- RUM Measure is $0.15 per 1,000 sessions on full traffic. Synthetic API tests are $5 per 10,000 runs, and browser tests are $12 per 1,000 runs.
How Dynatrace Charges
Dynatrace publishes one rate card and draws every capability from your annual commitment. Consume more than the commitment and you keep running at the same rates, billed monthly. Rates checked Nov 8, 2026:
- Foundation is $7 per host per month, Infrastructure Monitoring is $29, and Full-Stack Monitoring is $58 per 8 GiB host, billed at $0.01 per memory-GiB-hour. Full-Stack includes APM, code-level profiling, and Kubernetes Platform Monitoring.
- Log Analytics has two models. Pay-per-Query is $0.20 per GiB to ingest and process, $0.0007 per GiB-day to retain, and $0.0035 per GiB scanned. Bundled Queries is $0.20 per GiB ingested plus $0.02 per GiB-day, with queries included for 10 to 35 days.
- Metrics are $0.15 per 100,000 datapoints to ingest. Traces use the same $0.20 per GiB ingest rate and $0.0007 per GiB-day retention as logs.
- RUM is $2.25 per 1,000 sessions. Browser monitors are $4.50 per 1,000 synthetic actions, and HTTP monitors are $1.00 per 1,000 requests.
A Worked Example With 20 Hosts and 500 GB of Logs
Assume 20 hosts with 16 GiB of memory each, all running APM, plus 500 GB of logs a month at roughly 1 KB per event (500 million events), kept for 30 days. We treat GB and GiB as equal for simplicity, and the sums use only the list prices above.
- Hosts, Datadog: 20 × ($15 + $31) = $920 a month. Swap in APM Enterprise for profiling and it is 20 × ($15 + $40) = $1,100.
- Hosts, Dynatrace: a 16 GiB host is two 8 GiB units, so 20 × $116 = $2,320 a month. At 8 GiB per host the same fleet lists at 20 × $58 = $1,160. At 32 GiB it would be $4,640.
- Logs, Datadog, everything indexed for 30 days: $50 ingest + 500 × $2.50 = $1,300 a month. Index only 100 million events with exclusion filters and it falls to $50 + $250 = $300.
- Logs, Dynatrace, Pay-per-Query: $100 ingest + 500 × 30 × $0.0007 = $10.50 retention + 100 GiB scanned × $0.0035 = $0.35, which is about $110.85. With Bundled Queries it is $100 + 500 × 30 × $0.02 = $400.
Host memory decides who lists lower on hosts. Log shape decides who lists lower on logs. Dynatrace warns that processing can grow log volume by a factor of 2 or more, and filtering still counts toward Ingest and Process, so model your own volumes.
Agents, Data Models, and Query Languages
Dynatrace's pitch is automation. A single OneAgent per host discovers running processes and turns on instrumentation for what it finds, and Dynatrace Intelligence applies causation-based analysis over the Smartscape dependency graph and Grail. You do less manual wiring, and in exchange you work inside Dynatrace's model of your environment.
Datadog runs an open-source Agent and builds depth through integrations and per-product configuration. You choose which logs to index, which spans to retain, and which hosts get which tier. That control is why a Datadog bill can be tuned down, and why it needs tuning.
Querying differs too. Dynatrace stores logs, metrics, traces, and events in Grail and queries them with DQL, and every query against retained logs can generate scan charges unless you pick included queries. Datadog's Log Explorer searches indexed logs by facets and attributes, while Flex Logs adds paid compute for older data.
AI Assistants and Investigations
Both tools now sell autonomous investigations. Datadog's Bits Investigation forms hypotheses, gathers telemetry, and pinpoints likely root causes, billed in AI Credits. The published average is about 4 credits per investigation, so roughly $4 on an annual bundle ($1 per credit) or $5.20 on demand ($1.30 per credit). Watchdog, the built-in anomaly and root cause engine, needs no setup.
Dynatrace includes AI Exploration at no charge and meters Agentic Operations at $0.10 per AI unit. An autonomous multi-step investigation averages 30 units, or about $3, though Dynatrace is not billing those tasks inside Dynatrace Assist for a limited period. Treat all of these as averages that move with request complexity.
When to Pick Datadog
- You want a free tier or month-to-month billing before committing.
- Your hosts are large, so per-host pricing beats per-memory pricing.
- You want to pick products one at a time and tune each meter, including which logs get indexed.
- You have data residency needs that line up with its EU, Japan, Australia, or UK sites, or you want BYOC Logs.
- Your team already uses its integrations and wants the Agent plus OpenTelemetry side by side.
When to Pick Dynatrace
- You run many mid-sized or small hosts, where memory-based pricing stays low.
- You want automatic discovery and topology-based root cause analysis without hand-tuning each service.
- Your logs are large and rarely queried, so retention at $0.0007 per GiB-day and pay-per-scan suits you.
- You want one commitment and one rate card across infrastructure, logs, RUM, synthetics, and security.
- You need to run the platform yourself under Dynatrace Managed.
Where Sazabi Fits
Sazabi is AI-native observability and security that runs beside either tool. If you use an OpenTelemetry Collector, add an otlphttp exporter to send Sazabi a copy of your logs. If you use the Datadog Agent, additional_endpoints can dual-ship logs while Datadog keeps receiving them. Sazabi's agent investigates in Slack and opens issues, and its Hobby, Startup, and Growth plans include 200 GB of logs per 30-day cycle, then $0.10 per GB. Sazabi does not offer synthetic checks, so keep your synthetic monitoring tool for those.
Compare each tool one-on-one in Sazabi vs Datadog and Sazabi vs Dynatrace, browse the Datadog alternatives and Dynatrace alternatives, read about the OpenTelemetry integration, or check Sazabi pricing.
If Datadog is on your shortlist against other incumbents, see Datadog vs Splunk, Datadog vs New Relic, and Dynatrace vs Splunk.
Frequently Asked Questions
Is Dynatrace more expensive than Datadog?
It depends on host size and log volume. Dynatrace bills Full-Stack Monitoring at $0.01 per memory-GiB-hour ($58 a month per 8 GiB host), so cost scales with host memory. Datadog lists Infrastructure Pro at $15 and APM at $31 per host a month on annual terms regardless of memory. On logs the meters differ: Dynatrace charges $0.20 per GiB ingested plus retention and query fees, while Datadog charges $0.10 per GB ingested plus a per-event indexing fee. List prices checked Nov 8, 2026.
Does Datadog charge by host memory the way Dynatrace does?
The Datadog list prices for Infrastructure Pro and APM are per host, with no memory multiplier. Dynatrace Full-Stack Monitoring is metered per memory-GiB-hour, so a 16 GiB host lists at about twice the price of an 8 GiB host.
Can I send OpenTelemetry data to both Datadog and Dynatrace?
Yes. Datadog accepts OTLP through the DDOT Collector in the Datadog Agent, a standard OpenTelemetry Collector, or direct OTLP ingestion. Dynatrace accepts OTLP traces, metrics, and logs directly over OTLP/HTTP or through a Collector. A Collector with two exporters can feed both.
Can I run Dynatrace on my own infrastructure?
Dynatrace documents Dynatrace Managed, with its own deployment guides for servers, Kubernetes, and cloud platforms. Datadog is delivered from its hosted sites, with BYOC Logs as the option that keeps log storage in your own cloud object storage.
How are AI investigations billed on each tool?
Datadog bills Bits Investigation in AI Credits, averaging about 4 credits per autonomous investigation at $1 per credit on an annual bundle or $1.30 on demand. Dynatrace Agentic Operations costs $0.10 per AI unit, averaging 30 units per autonomous investigation, and those tasks are not billed inside Dynatrace Assist for a limited time. Both figures are published averages, not guarantees.
Sources
- Infrastructure Pro, APM, APM Enterprise, log ingestion and indexing, Flex Logs, RUM, synthetics, and AI Credit list prices (accessed Nov 8, 2026)
- Free tier (five Basic hosts, 1-day metric retention), APM allotments per host, AI Credit bundles and average credits per autonomous investigation (accessed Nov 8, 2026)
- The Datadog Agent is open source and runs on your hosts (accessed Nov 8, 2026)
- DDOT Collector in the Agent, OpenTelemetry Collector export, and direct OTLP ingestion (accessed Nov 8, 2026)
- Watchdog alerts, insights, and root cause analysis built in with no setup (accessed Nov 8, 2026)
- Flex Logs compute tiers, and that monitors and Watchdog are not supported in Flex Logs (accessed Nov 8, 2026)
- Datadog sites are independent and located in the US, EU, Japan, Australia, and the UK (accessed Nov 8, 2026)
- BYOC Logs runs in your own infrastructure with logs stored in your object storage (accessed Nov 8, 2026)
- Full-Stack Monitoring, Log Analytics, trace, RUM, synthetics, AI Exploration, and Agentic Operations list prices; DPS commitment model (accessed Nov 8, 2026)
- Ingest and Process, Retain, and Query dimensions, included queries, filtering behavior, and the 2x data volume warning (accessed Nov 8, 2026)
- OTLP ingestion of traces, metrics, and logs, directly or through a Collector, licensed under DPS (accessed Nov 8, 2026)
- One OneAgent per host and automatic instrumentation of discovered processes (accessed Nov 8, 2026)
- Grail is a single data lakehouse for logs, metrics, traces, and events (accessed Nov 8, 2026)
- Dynatrace Intelligence uses causation-based analysis over Grail and the Smartscape dependency graph (accessed Nov 8, 2026)
- Dynatrace Managed deployment documentation (accessed Nov 8, 2026)
- Sazabi plan prices, included usage, and per-GB overage rates (accessed Nov 8, 2026)
- Adding a Sazabi OTLP exporter beside existing Collector exporters (accessed Nov 8, 2026)
- Dual-shipping Datadog Agent logs to Sazabi with additional_endpoints (accessed Nov 8, 2026)