Datadog vs CloudWatch: Native AWS Monitoring or a Layer on Top of It
Prices and facts checked on vendor pages on Nov 8, 2026.
CloudWatch is the monitoring service built into AWS: metrics from AWS services arrive without setup, and you add logs, alarms, traces and dashboards on top. Datadog is a separate SaaS product that reads from AWS, including CloudWatch, and adds its own agents, APM and products. Many teams run both, so the real question is how much you want Datadog to do beyond what CloudWatch already holds.
The Short Answer
Pick CloudWatch if you run mostly on AWS, want metrics, logs and alarms with no extra vendor, and are happy to pay per GB and per metric. Pick Datadog if you want a richer experience across infrastructure, APM and security and will pay for it per host. If you pick Datadog on AWS, you are still using CloudWatch underneath, and some of the cost lands on your AWS bill.
Datadog vs CloudWatch at a Glance
How CloudWatch Charges
CloudWatch rates below are US East (N. Virginia) list prices from the pricing page, checked Nov 8, 2026. Other Regions differ.
- Logs: $0.50 per GB ingested on the Standard class, $0.25 on Infrequent Access. Storage is $0.03 per GB compressed, and Logs Insights queries are $0.005 per GB scanned. Live Tail is $0.01 per minute after 1,800 free minutes.
- Metrics: AWS services publish basic metrics for free. Custom metrics are $0.30 each for the first 10,000, $0.10 for the next 240,000, then $0.05 and $0.02.
- Alarms: $0.10 per standard-resolution alarm metric, $0.30 for high resolution, $0.50 per composite alarm. A log alarm is $0.10 plus $0.005 per GB scanned.
- API: GetMetricData is $0.01 per 1,000 metrics requested, which matters when a tool polls CloudWatch.
- Application monitoring: Application Signals is $1.50 per million for the first 100 million, and transaction search starts at $0.35 per GB ingested. X-Ray traces are $5 per million recorded.
How Datadog Charges
Datadog's list prices, on annual billing and checked Nov 8, 2026: Infrastructure Pro is $15 per host per month, APM is $31 per host, containers are $1 each, and custom metrics are $5 per 100. Logs are $0.10 per GB ingested plus $1.06, $1.27, $1.70 or $2.50 per million indexed events for 3, 7, 15 or 30 days. AI Credits are $500 per 500-credit bundle.
On AWS, Datadog's billing docs say only EC2 hosts, Lambda functions, CloudWatch custom metrics and containers are billable. Metrics for other AWS services, such as ELB, RDS and DynamoDB, do not incur Datadog charges. A host that runs the Agent and is also found by the AWS integration is not billed twice.
Two Worked Examples
Logs. Assume 250 GB per month at about 1 KB per event (250 million events), 15-day retention on Datadog, one month of compressed data stored on CloudWatch at the documented 0.15 compression ratio, and 50 GB scanned. The free tiers are ignored.
Custom metrics. Assume 20 EC2 hosts on Datadog Infrastructure Pro and 5,000 custom metrics. Pro includes 100 custom metrics per host, so 2,000 are covered and 3,000 are billed.
Each vendor defines a custom metric differently, so rerun this with your own counts. The Datadog figure also excludes the CloudWatch API calls Datadog makes for AWS service metrics.
How Datadog Reads CloudWatch
Datadog's AWS integration FAQ says it collects metrics through the CloudWatch GetMetricData endpoint every 10 minutes for each installed sub-integration, and that a large number of SQS, ELB, DynamoDB or custom-metric resources can affect your CloudWatch bill. For 2-3 minute latency it recommends CloudWatch Metric Streams with Amazon Data Firehose.
Logs follow a similar path. The Datadog Forwarder is a Lambda function that you subscribe to CloudWatch log groups, so you pay CloudWatch ingestion and then Datadog ingestion for the same line.
Logs, Alarms and APM
CloudWatch Logs organizes data in log groups that share retention settings and queries it with Logs Insights, with optional field indexes to cut scan volume. Alarms can watch metrics, PromQL queries, Logs Insights results or other alarms, and can trigger SNS, EC2 and Auto Scaling actions.
For APM, Application Signals supports Java, Python, Node.js and .NET on EKS, ECS and EC2, with SLOs and a service map. Datadog APM is a separate per-host product with indexed span retention you choose.
On AI, CloudWatch investigations starts from an alarm, a metric or a Logs Insights query. Datadog's Bits Investigation is an autonomous agent working from Datadog telemetry.
When to Pick Datadog
- You want host-level Agent data, not just the metrics AWS services publish.
- You want APM, infrastructure, security and incident tooling in one product.
- You want to keep indexed log volume under control with filters, not just retention settings.
When to Pick CloudWatch
- Your workloads are mostly on AWS and you want alarms wired to AWS actions.
- You want one AWS bill, with IAM and cross-account observability you already manage.
- Your log volume is moderate and per-GB pricing is easier to predict than per-event.
- You do not want another agent or vendor in the data path.
Where Sazabi Fits
Sazabi is AI-native observability and security that works beside both tools. When you connect an AWS account, it installs subscription filters on the CloudWatch log groups you select, so logs reach Sazabi without changing what CloudWatch or Datadog receive. Its agent investigates in Slack and opens issues. For details, see Sazabi vs CloudWatch, Sazabi vs Datadog, CloudWatch alternatives and the AWS integration.
Frequently Asked Questions
Does Datadog replace CloudWatch?
Not on AWS. Datadog reads AWS service metrics through the CloudWatch GetMetricData API, so CloudWatch keeps producing and storing them. Datadog polls each AWS integration every 10 minutes by default, or about 2-3 minutes if you use CloudWatch Metric Streams with Amazon Data Firehose.
Is CloudWatch cheaper than Datadog for logs?
At list prices CloudWatch charges $0.50 per GB ingested and no per-event indexing fee, while Datadog charges $0.10 per GB ingested plus $1.06 to $2.50 per million indexed events on annual billing. At 250 GB a month with every event indexed for 15 days, the arithmetic on this page gives about $126 for CloudWatch and $450 for Datadog. Indexing only part of your logs in Datadog lowers its total.
Does running Datadog increase my CloudWatch bill?
It can. Datadog calls the CloudWatch API to collect metrics, and its docs warn that a large number of resources in an integration such as SQS, ELB or DynamoDB can affect your CloudWatch bill. CloudWatch lists GetMetricData at $0.01 per 1,000 metrics requested.
How do CloudWatch logs get into Datadog?
The usual route is the Datadog Forwarder, an AWS Lambda function you subscribe to CloudWatch log groups. It also forwards S3 logs and SNS and Kinesis events. The CloudWatch ingestion charge still applies, because the logs are stored in CloudWatch before they are forwarded.
Sources
- Infrastructure Pro, APM, container, custom metric, log ingestion, indexed log and AI Credit list prices (accessed Nov 8, 2026)
- Free tier of up to five Basic hosts with 1-day metric retention (accessed Nov 8, 2026)
- Datadog gathers CloudWatch metrics through GetMetricData every 10 minutes, which can affect your CloudWatch bill; Metric Streams with Amazon Data Firehose give 2-3 minute latency (accessed Nov 8, 2026)
- Only EC2 hosts, Lambda functions, CloudWatch custom metrics and containers are billable by Datadog; no double billing for EC2 hosts running the Agent (accessed Nov 8, 2026)
- Pro accounts get an allotment of 100 indexed custom metrics per host (accessed Nov 8, 2026)
- The Datadog Forwarder is a Lambda function that ships CloudWatch and S3 logs, and SNS and Kinesis events, to Datadog (accessed Nov 8, 2026)
- Bits Investigation is described as an autonomous AI agent that investigates production issues (accessed Nov 8, 2026)
- US East (N. Virginia) rates for log ingestion and storage classes, Logs Insights queries, Live Tail, custom metrics tiers, alarms, API requests, Application Signals, X-Ray and free tier (accessed Nov 8, 2026)
- CloudWatch product overview: metrics, logs and traces across your AWS environment, composite alarms, anomaly detection and cross-account observability (accessed Nov 8, 2026)
- CloudWatch Logs classes, Logs Insights query language, field indexes and Live Tail (accessed Nov 8, 2026)
- Log groups share retention settings; metric filters turn log events into metrics (accessed Nov 8, 2026)
- Metric, PromQL, log and composite alarms and the actions they can trigger (accessed Nov 8, 2026)
- Application Signals supports Java, Python, Node.js and .NET on EKS, ECS and EC2, with SLOs and a service topology map (accessed Nov 8, 2026)
- CloudWatch OTLP endpoints for logs, with SigV4 or bearer token authentication (accessed Nov 8, 2026)
- CloudWatch investigations is a generative AI assistant that surfaces telemetry and root-cause hypotheses from alarms, metrics or Logs Insights queries (accessed Nov 8, 2026)
- Sazabi installs subscription filters on selected CloudWatch log groups through a cross-account IAM role (accessed Nov 8, 2026)