Grafana Cloud synthetic monitoring pricing and cost per test execution
Grafana meters synthetics by the execution, and probes, checks, run time and interval all multiply together. Put your real configuration in and the calculator runs Grafana's own formula, then prices it.
Grafana Cloud synthetic monitoring cost calculator
LiveRates read in September 2026 from grafana.com/pricing, USD list, using Grafana's published formula probes x tests x duration x (43,200 / frequency) with run time rounded up to the whole minute. Volume bands are applied marginally. The $19 Cloud Pro platform fee is included; metrics, logs, traces and per user charges are not.
How much does Grafana Cloud synthetic monitoring cost?
Grafana Cloud bills synthetic monitoring by the test execution, which is one test running in one probe location for one minute of run time, rounded up. API test executions cost $5.00 per 10,000 and browser test executions cost $50.00 per 10,000, on top of a $19 a month Cloud Pro platform fee. Both the free plan and the $19 Pro plan include exactly 100,000 API and 10,000 browser executions a month, so the platform fee buys no extra capacity at all, only the right to exceed the cap and pay for the overage. Because probes, tests, duration and frequency all multiply together, one HTTP check running every minute from five probes consumes 216,000 executions and costs about $77 a month by itself.
The published rate card
- Cloud Pro platform fee
- $19.00 / mo
- API executions
- $5.00 / 10k
- Browser executions
- $50.00 / 10k
- Included, free plan
- 100k + 10k
- Included, Pro plan
- 100k + 10k
- Enterprise minimum
- $25,000 / yr
The $19 plan includes exactly the same synthetic allowance as the free one
Read Grafana's two plan cards next to each other and the free one says it is limited to 100,000 API test executions and 10,000 browser test executions per month. The Pro card says the platform fee of $19 per month includes 100,000 API test executions and 10,000 browser test executions per month, then pay as you go. Those are the same two numbers.
So the $19 is not buying synthetic capacity. It buys 13 months of retention instead of 14 days, 8x5 email support instead of the community forum, and, the part that actually matters here, permission to go over the line. On the free plan the allowance is a ceiling and checks stop when you reach it. On Pro it is a floor, and everything above it meters. That distinction is worth paying for if you are running anything production shaped, but it is worth knowing that you are paying for the meter rather than for headroom.
The practical consequence is that there is no middle step. You are either inside a fixed allowance that costs nothing, or you are on a variable bill with no cap on it. Teams that want a predictable number in the budget line tend to find that uncomfortable, which is the whole reason flat rate monitoring vendors exist.
Included synthetic executions per month
| Plan | API | Browser |
|---|---|---|
| Free | 100,000 | 10,000 |
| Pro, $19 / mo | 100,000 | 10,000 |
| Difference | 0 | 0 |
Free is capped at that amount. Pro adds pay as you go above it at $5.00 and $50.00 per 10,000.
The browser ladder is the API ladder multiplied by ten, at every single step
Most vendors price their browser and API products independently, and the gap between them lands wherever it lands. Grafana did something cleaner, and as far as we can tell nobody has written it down. Take every number on the API ladder and multiply it by ten, and you have the browser ladder exactly: the three rates, and the two volume thresholds that separate them.
| Measure | API testing | Browser testing | Ratio |
|---|---|---|---|
| Entry rate per 10,000 | $5.00 | $50.00 | 10.0x |
| Second band rate | $4.45 | $44.50 | 10.0x |
| Third band rate | $4.15 | $41.50 | 10.0x |
| Second band starts at | 1,000,000 executions | 100,000 executions | 10.0x |
| Third band starts at | 2,500,000 executions | 250,000 executions | 10.0x |
| Second band starts at, in dollars | $500 | $500 | Identical |
| Third band starts at, in dollars | $1,250 | $1,250 | Identical |
| Included free each month | 100,000 executions | 10,000 executions | 10.0x |
| Value of that inclusion | $50.00 | $50.00 | Identical |
What that symmetry actually means for your bill
Because both the price and the threshold scale by the same factor, the volume discount is not really triggered by executions at all. It is triggered by money. One million API executions at $5.00 per 10,000 is $500 of spend. One hundred thousand browser executions at $50.00 per 10,000 is also $500 of spend. The third band opens at $1,250 on both. So whichever product you run, you reach Grafana's first discount at $500 a month and its deepest published discount at $1,250 a month.
And the free allowance is worth the same either way
The same arithmetic runs through the free tier. The 100,000 included API executions are worth $50.00 at the entry rate. The 10,000 included browser executions are worth $50.00 at the entry rate. Grafana gives every account exactly $50 a month of synthetic monitoring, then lets you decide whether to spend it on a lot of cheap checks or a few expensive ones. That is a more coherent free tier than most of this market ships, and it deserves the credit.
One caveat worth stating plainly. Grafana describes these as automatic volume discounts and says the effective rate drops as consumption rises, but it does not say whether a band rate applies only to the executions inside that band or to your whole volume once you cross it. This page models them marginally, which is the more conservative reading and matches how the rest of the market publishes ladders. If Grafana applies them flat, real bills at high volume are lower than the figures here. We are flagging that rather than picking the convenient answer quietly.
A check that slows from 59 seconds to 61 seconds doubles its own bill
The billing unit is a test running in a probe location per minute of run time, and Grafana rounds that run time up to the whole minute before counting. Nothing in that sentence looks dangerous until you notice that duration is a multiplier in the formula rather than a rounding detail at the end of it.
So the cost of a check is a step function of how long it takes. Everything from one second to sixty seconds costs the same. At sixty one seconds the cost doubles, and it stays doubled until two minutes, where it goes up by half again. A single API check running every minute from five probes costs $77 a month at $19 plus overage while it completes inside a minute. Let it drift to 61 seconds and the same check costs $185, an increase of $108 a month for two seconds of extra latency.
This matters most for browser journeys, which are the ones that naturally sit near the minute line. A login, a search and a checkout in one script is comfortably a 45 to 90 second job, which means the same test can land on either side of the cliff depending on how the site is performing that week. The month your application gets slower is the month your monitoring bill doubles, which is an unfortunate time for it to happen.
One browser journey, 3 probes, one run
| Actual run time | Billed as | Executions | Cost per run |
|---|---|---|---|
| Completes in 45 seconds | 1 minute | 3 | $0.015 |
| Completes in 59 seconds | 1 minute | 3 | $0.015 |
| Completes in 61 seconds | 2 minutes | 6 | $0.030 |
| Completes in 1 min 59 sec | 2 minutes | 6 | $0.030 |
| Completes in 2 min 1 sec | 3 minutes | 9 | $0.045 |
At the $50.00 per 10,000 entry rate. Grafana's own worked example uses a 1.5 minute test in 5 locations and counts it as 10 executions, which is the same round-up.
Where the free 100,000 executions actually runs out
A hundred thousand sounds like a lot until you divide it by the number of minutes in a month. The useful unit here is the probe-and-check pair, meaning one check running from one location. Each pair consumes 43,200 executions a month at a one minute interval, so the free allowance holds two of them. At five minutes it holds eleven. The table below is the whole story of who pays nothing and who gets a bill.
| Interval | Executions per pair, per month | Pairs inside the free 100,000 | What that looks like |
|---|---|---|---|
| Every 30 seconds | 86,400 | 1 pair | One check from one probe already uses 86,400 of the 100,000 |
| Every minute | 43,200 | 2 pairs | Two probes on one check fits. Three probes does not |
| Every 2 minutes | 21,600 | 4 pairs | Four checks from one probe, or two checks from two |
| Every 5 minutes | 8,640 | 11 pairs | The usual sweet spot for a small estate |
| Every 10 minutes | 4,320 | 23 pairs | Comfortable, but a 10 minute gap is a long outage |
| Every 15 minutes | 2,880 | 34 pairs | Plenty of room, very little urgency |
Read that column on the right and the shape of Grafana's synthetic product becomes clear. It is generous for a handful of endpoints checked at a relaxed cadence, and it stops being generous the moment you want either speed or geographic coverage. Those are precisely the two things uptime monitoring is about, which is why teams who arrive here looking for an uptime tool tend to leave with a larger number than they expected.
Six real configurations, priced end to end
Each row runs Grafana's formula, subtracts the free 100,000 API executions, applies the rate ladder marginally and adds the $19 platform fee. The last column is what the same estate costs on our own flat plans, where the interval and the region count are plan features rather than meters. The first two rows are the ones worth pausing on, because Grafana wins them outright.
| Configuration | Executions / mo | Metered overage | Grafana total / mo | Uptimehub / mo |
|---|---|---|---|---|
| 5 API checks, every 5 min, 2 probes | 86,400 | Free | $19.00 | $12.00 |
| 10 API checks, every 5 min, 1 probe | 86,400 | Free | $19.00 | $12.00 |
| 1 API check, every minute, 5 probes | 216,000 | $58.00 | $77.00 | $12.00 |
| 20 API checks, every minute, 3 probes | 2,592,000 | $1,163.94 | $1,182.94 | $12.00 |
| 100 API checks, every minute, 6 probes | 25,920,000 | $10,845.30 | $10,864.30 | $39.00 |
| 100 API checks, every 30 sec, 6 probes | 51,840,000 | $21,602.10 | $21,621.10 | $39.00 |
The jump between row three and row four is the part worth understanding, because it is not a pricing decision, it is a multiplication. Going from one check to twenty multiplies by twenty. Nothing about the rate card changed. This is what people mean when they say consumption pricing is hard to budget: the bill is a product of four numbers you adjust for operational reasons, and none of those adjustments feels like a purchase at the time you make it.
When Grafana Cloud is the right answer, and when it is the wrong shape
Grafana is the better buy when
- You already run Grafana Cloud and want check results in the same dashboards as your metrics, logs and traces, correlated on one timeline.
- Your estate is small and your cadence is relaxed. Eleven probe-and-check pairs at five minutes genuinely costs nothing, and the free tier is not a trial.
- You want scripted multi-step API flows and k6 load testing from the same platform and the same account.
- You have engineers who are comfortable owning a consumption bill and tuning it, and the cost lands in a budget that already has Grafana in it.
A dedicated monitor is the better buy when
- You want minute or sub-minute checks from several regions, which is exactly the configuration the execution meter punishes hardest.
- You need a number you can put in a budget and defend, rather than one that moves when an engineer adds a probe.
- You want a branded status page for customers, which Grafana's synthetic product does not provide.
- The people who need to see it are not Grafana users, and buying them seats at $8 each changes the arithmetic again.
Both of these can be true at once, and for a lot of teams they are. Synthetic checks inside Grafana for the deep, scripted, correlate-with-everything work, and a flat rate monitor for the boring question of whether the site is up, checked often, from everywhere, with a status page attached. The mistake is using the consumption meter for the boring question, because that is the one you want running constantly and from many places, and those are the two multipliers.
Where these numbers come from
Every figure was read in September 2026 from grafana.com/pricing fetched with a browser user agent, plus Grafana's synthetic monitoring invoice documentation for the billing unit. The pricing page renders each metered product in its own block with a stable identifier, so the API and browser ladders were parsed separately by block rather than pulled out of flattened page text. That matters more than it sounds: flattened, the two ladders become one run of six numbers, and transposing them would put browser checks at a tenth of their real price.
The model was then validated against Grafana's own worked example before any of it was used. Grafana states that a test running in 5 locations taking 1.5 minutes consumes 10 executions per run. Our implementation returns 10 for that input, which confirms both the multiplication and the round-up to the whole minute. We do not publish an Enterprise unit rate, because Grafana publishes only a $25,000 a year minimum commit, and we do not publish the active series credits that apply to accounts created before 13 February 2026, because they depend on your signup date rather than on the rate card.
Grafana Cloud synthetic monitoring pricing questions
What is a test execution?
A test execution is one synthetic test running in one probe location for one minute of run time. Grafana rounds the run time up to the whole minute before counting, so a test that finishes in 20 seconds and one that finishes in 59 seconds both count as one execution per probe, and a test that takes 61 seconds counts as two. Grafana's own example is a test running in 5 locations that takes 1.5 minutes, which consumes 10 executions per run.
How do I estimate the number of test executions?
Grafana publishes the formula: probes x tests x duration x (43,200 / frequency), where 43,200 is the number of minutes in a 30 day month, duration is the run time rounded up to the nearest minute and frequency is the check interval in minutes. Every term multiplies, so doubling the probe count or halving the interval doubles the bill. The calculator at the top of this page runs that formula and then applies the published rate ladder.
How much does Grafana Cloud synthetic monitoring cost?
API test executions cost $5.00 per 10,000 and browser test executions cost $50.00 per 10,000, with volume discounts to $4.45 and $4.15 for API and $44.50 and $41.50 for browser. On top of that sits the Grafana Cloud Pro platform fee of $19 a month. The free plan and the Pro plan both include 100,000 API and 10,000 browser executions a month, so a small estate at a 5 minute interval genuinely pays nothing for synthetics.
Is Grafana synthetic monitoring free?
Yes, up to 100,000 API and 10,000 browser test executions a month, and that allowance is real rather than a trial. It holds roughly eleven probe-and-check pairs at a five minute interval. It stops holding at higher frequency: a single check running every minute from three probes consumes 129,600 executions and is already over the cap. On the free plan the cap is hard, so checks stop rather than bill.
Does the $19 Pro plan include more synthetic checks than the free plan?
No. Grafana lists the same 100,000 API and 10,000 browser test executions on both. The difference is what happens at the ceiling: the free plan is described as limited to that amount, while Pro adds pay as you go above it. So the $19 platform fee buys 13 months of retention, email support and the right to exceed the cap, but not one extra execution of included capacity.
Why is browser testing ten times the price of API testing?
Browser executions run a real browser rather than an HTTP request, so they cost Grafana more to run. The interesting part is how exactly the factor of ten holds: every rate and every volume threshold on the browser ladder is precisely ten times its API counterpart. That means both products reach their volume discounts at the identical dollar spend, $500 and $1,250 a month, and both free allowances are worth exactly $50.
What counts as a probe location?
A probe is a geographic location Grafana runs the check from, and each one multiplies your execution count directly. Two probes cost twice as much as one for the same check at the same interval. This is the term teams most often forget when they estimate, because adding a region feels like a reliability decision rather than a billing decision, and in this model it is both.
Does Grafana Cloud offer volume discounts?
Yes, and they are published rather than negotiated. API drops from $5.00 to $4.45 per 10,000 above one million executions and to $4.15 above 2.5 million, an 11% and then a 17% reduction. Browser drops from $50.00 to $44.50 and $41.50 at one tenth those volumes. Below those thresholds the entry rate is what you pay, and reaching the deepest published band takes $1,250 a month of synthetic spend.
What is the minimum for Grafana Cloud Enterprise?
Grafana publishes a minimum commit of $25,000 per year for Enterprise, which is the only Enterprise figure on the pricing page. Unit rates at that tier are described as a scalable unit price based on annual commit and no numbers are given, so we do not model Enterprise here. That minimum is about 110 times the annual cost of the $19 Pro platform fee.
Is Grafana Cloud cheaper than a dedicated uptime monitoring tool?
For a small estate at a relaxed interval, yes, because the free allowance covers it outright. For uptime monitoring shaped work it stops being cheaper quickly, since frequency and probe count both multiply the bill rather than being plan features. Twenty checks every minute from three probes is 2.59 million executions a month, around $1,183, where flat rate monitoring tools price the same estate in the tens of dollars.
Checking more often should not cost more
Uptimehub checks your sites, APIs and ports every 30 seconds from six regions for a flat monthly price, with the interval and the region count included rather than metered. Branded status pages are on every plan, TLS and domain expiry are watched alongside the endpoint, and the bill is the same number every month whether your checks take one second or ninety.
Related breakdowns
Synthetic monitoring pricing compared
Datadog, New Relic, Dynatrace, Grafana, AWS and Checkly priced per check run in one table, because every vendor meters a different unit.
Sentry uptime monitoring pricing
The other vendor whose paid plan includes no more monitoring capacity than its free one, at $1.00 per monitor after the first.
Uptime monitoring pricing compared
Every flat rate vendor in one table, sorted by billing model, because the model decides the bill more than the sticker does.