Datadog Synthetics Pricing: What API and Browser Tests Cost

Datadog Synthetic Monitoring is billed per test run, and the unit is different for the two test types. API tests are $5 per 10,000 test runs per month on annual billing. Browser tests are $12 per 1,000 test runs per month. That works out to $0.0005 per API run and $0.012 per browser run, so a browser run costs 24 times what an API run costs. Because you are billed per run rather than per monitor, your bill is set by three numbers you choose: how many tests you have, how many locations each one runs from, and how often it fires.

The confusing part of pricing Datadog Synthetics is not the rate card. It is that the rate card does not answer the question people actually have, which is what their setup will cost. A per-monitor tool lets you count monitors and stop. Here you have to do a multiplication first, and the number that comes out surprises people, usually because of the check interval rather than the test count.

How much does Datadog Synthetics cost?

These are the list prices from Datadog's own pricing comparison page, read in August 2026. Datadog publishes three rates for each unit depending on how you commit.

Datadog's documentation defines the two units precisely. An API test is "an HTTP or HTTPS request against a single URL". A browser test is "a simulation of a scripted sequence of user actions on a web-based application using a virtualized web browser", and a single browser test run covers a web transaction of up to 25 steps. Vendor pricing moves, so confirm the current figures on Datadog's page before you budget against them.

Read those two rows carefully, because the trap is in the middle column. API tests are quoted per ten thousand runs and browser tests per one thousand. The prices look like $5 against $12, a bit over double. Per actual run they are $0.0005 against $0.012, which is 24 times more expensive. Any estimate built by comparing $5 to $12 will be wrong by more than an order of magnitude on the browser side.

What does a real monitoring setup cost?

Your monthly run count is the number of tests multiplied by the number of locations each test runs from, multiplied by how many times it fires in a month. A 30 day month contains 43,200 minutes, so a test on a one minute interval fires 43,200 times, on five minutes 8,640 times, on fifteen minutes 2,880 times, and hourly 720 times. Those are per test, per location.

Here is what that produces at the annual rate for a small API setup, ten endpoints checked from three locations.

Ten endpoints is a modest setup. The same ten endpoints cost $648 a month or $10.80 a month depending only on a dropdown, and the tests are otherwise identical. That is a 60 fold spread on a single setting, and it is the single most important thing to understand before you commit to a plan.

Browser tests are where budgets actually break, because the 24 times per run difference compounds against the same multipliers.

One browser test, from one location, every five minutes, is $103.68 a month. That is a single scripted journey watched at a normal interval, and it costs more than most teams pay for their entire uptime monitoring stack. Five journeys across three locations at the same interval is $1,555.20 a month. Neither number appears anywhere on a pricing page, and both are just the rate card multiplied out.

Why the check interval is the price

Under per-run billing, halving the interval doubles the bill, exactly and permanently. Moving those ten API endpoints from a one minute to a five minute interval takes the monthly cost from $648 to $129.60, a saving of $518.40, or 80%. Nothing about the tests changed. You are simply buying fewer executions.

The catch is that the interval is not a free variable, because it sets how fast you find out about an outage. Worst case detection delay is one full check interval and the average is half of one, so moving from one minute to five minutes to save that $518.40 also means an outage can run for up to five minutes before anything notices. Against a 99.9% monthly target, which allows 43 minutes 12 seconds of downtime, a single worst case five minute detection has spent about 12% of the entire month's budget before anyone was paged. We work through how to pick that number against a target in how often you should monitor a website.

That is the genuine tension in consumption pricing, and it is worth naming plainly: the setting that controls your bill is the same setting that controls your detection speed, so cost control and reliability pull directly against each other. Under flat per-monitor pricing they do not, because the interval is included and choosing a faster one costs nothing.

The on demand penalty

The third column of the rate card is the one that catches teams that grow mid-contract. Usage beyond your commitment is billed on demand at $7.20 per 10,000 API runs and $18 per 1,000 browser runs, against annual rates of $5 and $12. That is 44% more for API runs and 50% more for browser runs.

The practical risk is that run volume can jump without anyone making a purchasing decision. Adding two locations to an existing test raises its run count by the same proportion as adding two more tests. Tightening one interval from five minutes to one multiplies that test's runs by five. Both are one click in the UI, neither looks like a spending decision, and both land at the on demand rate if you are already at your commitment. If you run consumption priced tooling across several vendors, this is the category of spend worth watching centrally rather than per tool, which is the argument for tracking cloud and SaaS spend in one place instead of discovering it on an invoice.

How to check what you are actually using

Do not estimate this from memory. Datadog exposes usage as metrics you can graph directly, which turns the whole question into a dashboard:

• sum:datadog.estimated_usage.synthetics.api_test_runs{*}.as_count()
• sum:datadog.estimated_usage.synthetics.browser_test_runs{*}.as_count()

Graph both as a monthly sum, divide the API figure by 10,000 and the browser figure by 1,000, then multiply by your rates. That gives you the real number rather than the one you assumed, and it catches the two things people always miss: tests someone added and forgot, and tests running from more locations than anyone intended. It is also the fastest way to find out whether your locations multiply your run count the way you think they do, since your own account is the authoritative answer to that.

Three ways monitoring vendors meter the same job

Datadog's model is one of three you will meet while shopping, and the differences matter more than the headline prices, because they decide which of your choices are free and which cost money.

Credit based pricing is its own puzzle and we have taken it apart separately in our Uptrends alternative and credit model breakdown, where an uptime monitor costs one credit no matter how often it runs. For a wider view of how ten vendors meter this same job, the uptime monitoring pricing comparison lays the billing models side by side.

For context on scale, our own Starter plan is $9 a month billed yearly for 20 monitors at one minute checks from six regions, with a branded status page included. A single Datadog browser test at five minutes is $103.68. Those are not competing products, and that is the actual point: they are priced for different jobs, and picking on price alone gets you the wrong one.

When Datadog Synthetics is worth the money

Consumption pricing is not a trick, and there are teams for whom it is clearly the right buy. Three cases stand out.

The first is when you are already deep in Datadog. If your logs, APM traces and infrastructure metrics live there, a synthetic failure that links straight into the trace of the request that failed is worth real money in time saved during an incident. No standalone uptime tool can offer that, and splitting your data to save a subscription is usually a bad trade.

The second is scripted browser journeys. If a broken checkout has to page someone before a customer finds it, you need a tool that drives a real browser through the steps and fails when step four breaks. That is what browser tests are for, it is genuinely expensive to run, and the $0.012 per run is buying real compute rather than margin.

The third is bursty or CI driven testing. Per run billing is the fairest model that exists for tests that run only when a pipeline runs. Paying per monitor for something that executes forty times a month is worse value than paying for forty runs.

What consumption pricing suits badly is the ordinary case: a fixed set of endpoints you want watched continuously and quickly, forever. There the run count is large, predictable and permanent, the interval you actually want is the fastest one available, and every one of those properties makes per run billing expensive. If that is your situation, a Datadog Synthetics alternative priced per monitor will be cheaper and easier to forecast, and for API endpoints specifically the best API monitoring tools roundup compares what each option actually checks. Datadog is also not the only vendor metering you this way, and the rates are not close: New Relic charges ten times more per check and Checkly half as much, which we set side by side in the synthetic monitoring pricing comparison.

Is Datadog Synthetics expensive?

It depends entirely on your interval, and that is not a dodge. Ten API endpoints from three locations cost $10.80 a month at hourly checks and $648 a month at one minute checks. The same tool is cheap or expensive based on one setting. Browser tests are expensive by any reading at $0.012 per run, though they are also doing far more work per run than an HTTP request. The honest summary is that Datadog Synthetics is priced fairly for what it does and priced badly for continuous high frequency availability checking, which is the job most people are shopping for when they land on a synthetics pricing page.

Does Datadog Synthetics have a free tier?

Datadog offers a free trial of the platform rather than a permanently free synthetics allowance, and synthetic monitoring is sold as usage on top of your Datadog subscription rather than as a standalone product with its own free plan. If a free tier is what you are looking for, you are comparing against tools built to be bought on their own, and the comparison to run is total monthly cost at the interval you actually intend to use, not the entry price.