GCP uptime check pricing: Google Cloud uptime checks and GCP uptime monitoring, priced per region
Short answer: Google Cloud Monitoring charges $0.30 per 1,000 uptime check executions, with 1 million executions free per project per month. That free allowance sounds enormous and it is not, because Google bills every region separately. A check running from the three USA checkers at the default 60 second period is 131,499 executions a month, so the free million covers 7.6 URLs, not thousands. Past that you pay $39.45 per URL per month. The floor is also hard: 60 seconds is the fastest period Google offers at any price. Both numbers are worked through below.
Last updated September 2026 · US dollar list prices, cloud.google.com/stackdriver/pricing
alertping ▸ run check
live
▸ type a domain and run a real-feel check
▸ probes from 3 regions · FRA · IAD · SIN
▸ waiting…▌
queued probing▌
If ever goes down, you get:
Alert fired ▸ 2 channels · 6.2 s after first failure
AlertPing app
● DOWN : HTTP timeout confirmed from 3/3 regions (FRA, IAD, SIN). Incident opened.
sms · on-call
AlertPing: DOWN. Confirmed 3/3 regions . First fail: Frankfurt.
read this first
The free million executions covers 7.6 URLs
Almost every write up of GCP uptime check pricing quotes two numbers and stops: $0.30 per 1,000 executions, 1 million executions free per project each month. Both are correct. Together they give a badly wrong impression, because they leave out the sentence that does the real work in Google's own pricing page: a check that executes in three regions counts as three executions.
That single line changes the arithmetic by a factor of three to six, and it is not an edge case you can design around. Google's own uptime check documentation tells you to identify at least three checkers, because a single vantage point cannot distinguish a dead service from a bad route. The USA region alone contains three checkers, in Oregon, Iowa and Virginia. Europe, South America and Asia Pacific hold one each, and selecting Global issues the check from every uptime check region. So the smallest configuration a sane engineer would ship is already billing three executions per period, and the default Global setup bills six.
Run that through a month. A 60 second period is 43,833 checks. Times three USA checkers is 131,499 executions for one URL. The free million divided by 131,499 is 7.60. Your eighth production hostname is where the invoice starts, and if you left the check on Global it is your fourth.
The multiplier
Regions, not checks
The billing unit is a regional execution, not a check. One URL, one minute, three USA checkers is three billable events every 60 seconds. Nothing in the console labels this, and the cost only becomes visible after the free allowance runs out, which is usually a month or two after somebody added the checks.
The allowance
Per project, not per account
The 1 million executions are free per Google Cloud project. Executions are charged to the billing account in which they are defined. Teams that split environments across many projects get many allowances, and teams that centralize monitoring into one observability project get exactly one. The tidier architecture is the more expensive one here.
The ceiling
100 checks per metrics scope
Cloud Monitoring caps uptime checks at 100 per metrics scope, and a public uptime check accepts a maximum of 3 ICMP pings. If you need more than 100 monitored endpoints you are splitting scopes, which splits the dashboards and the alerting policies along with them.
the bill
What Google Cloud uptime monitoring costs per URL
Every figure below is (43,833 / period in minutes) x checkers x $0.0003 per execution, before the free allowance is applied. The last column is the free allowance expressed the way it actually matters: how many URLs you can run in that configuration before the meter starts.
| Configuration, one URL | Executions per month | Cost per month, unmetered | URLs inside the free million |
|---|---|---|---|
| 15 minute period, 3 USA checkers | 8,767 | $2.63 | 114.07 |
| 10 minute period, 3 USA checkers | 13,150 | $3.94 | 76.05 |
| 5 minute period, 3 USA checkers | 26,300 | $7.89 | 38.02 |
| 60 second period, 3 USA checkers (the default period) | 131,499 | $39.45 | 7.60 |
| 60 second period, Global (6 checkers) | 262,998 | $78.90 | 3.80 |
| Synthetic monitor, 60 second period (billed separately at $1.20 per 1,000) | 43,833 | $52.60 | 100 runs free |
Uptime check rate $0.30 per 1,000 executions with 1 million free per project, in effect since October 1, 2022. Synthetic monitors are a separate SKU at $1.20 per 1,000 executions with a free allowance of 100 executions per billing account, in effect since November 1, 2023. Both read from cloud.google.com/stackdriver/pricing in September 2026. A 30.44 day month is 43,833 minutes.
| Fleet size at 60 second period, 3 USA checkers | Billable executions | Google Cloud per month | AlertPing per month |
|---|---|---|---|
| 5 URLs | 0 | $0.00 | $19 |
| 10 URLs | 314,990 | $94.50 | $19 |
| 20 URLs (our Starter monitor count) | 1,629,980 | $488.99 | $19 |
| 100 URLs (our Team monitor count, and Google's per scope cap) | 12,149,900 | $3,644.97 | $59 |
Billable executions are gross executions minus the 1 million free allowance. AlertPing Starter is $19 a month for 20 monitors at 60 second checks; Team is $59 a month for 100 monitors, and 30 second checks start there.
The three break-even points
At a 60 second period across three USA checkers, Google Cloud's uptime check bill passes our $19 Starter plan at 8.09 URLs, our $59 Team plan at 9.10 URLs, and our $189 Business plan at 12.40 URLs.
Those three numbers sit within four URLs of each other, which is the part worth internalizing. Once the free allowance is gone the meter climbs so steeply that the distance between our cheapest plan and our third plan is about four hostnames of Google Cloud usage. At 100 URLs the gap is $3,644.97 against $59, and the $59 plan is checking twice as often.
Where Google Cloud genuinely wins
Below roughly seven URLs at a 60 second period, GCP uptime monitoring costs you nothing at all and we cost $19. That is not close. If you run a handful of endpoints in one project, use the free allowance and spend the money elsewhere. At a 15 minute period the free tier stretches to 114 URLs, which covers most companies outright.
The other honest point: Google's 60 second floor matches our Starter plan exactly. We are not faster than Cloud Monitoring until you reach our $59 Team plan. And a check result that lands in the same Cloud Monitoring workspace as your application metrics can be correlated against the request that failed, in one query, which no external service can do for you.
the hard floor
60 seconds is the fastest period Google sells
The uptime check API is explicit, and it is a short list. The period field accepts
60s, 300s,
600s and 900s.
One minute, five minutes, ten minutes, fifteen minutes. Nothing else. The default is 60s and
there is no faster tier, no enterprise override and no support ticket that changes it. Timeout
is separately configurable between 1 and 60 seconds.
That matters because a single missed check is the smallest outage you can record. If your check runs every 60 seconds, you cannot evidence an outage shorter than 60 seconds, and every outage you do record is rounded to the nearest whole check. Set that against the monthly downtime budget your contract actually specifies and the picture gets uncomfortable quickly.
| Availability target | Monthly downtime budget | One missed check at 60s | One missed check at 5 min | One missed check at 15 min |
|---|---|---|---|---|
| 99.9% | 43.83 minutes | 2% of budget | 11% of budget | 34% of budget |
| 99.95% | 21.92 minutes | 5% of budget | 23% of budget | 68% of budget |
| 99.99% | 4.38 minutes | 23% of budget | 114% of budget | 342% of budget |
Budgets computed against a 43,833 minute month. A figure above 100% means a single missed check already exceeds the entire monthly allowance, so the target cannot be evidenced at that period.
The reading you can actually report
Below 100%, the best number a check can produce is one missed check. At a 60 second period that is 99.99772%. At five minutes it is 99.98859%, which is already below a 99.99% commitment. At fifteen minutes it is 99.96578%, below 99.99% and below 99.95% too.
So if somebody moved your checks to a 15 minute period to stay inside the free million, they also quietly made your 99.95% target unprovable. That trade is real and almost nobody makes it deliberately.
Measuring Google from inside Google
The second limitation has nothing to do with price. Cloud Monitoring uptime checks run on Google Cloud infrastructure. When the failure is Google Cloud, your evidence of the failure is produced by the thing that failed, and the alerting policy that should have paged you depends on the same platform.
This is not a knock on Google's reliability. It is the reason auditors and enterprise customers ask for a measurement taken somewhere else. We check from three regions, Frankfurt, Virginia and Singapore, and confirm a failure in all three before it counts as downtime, which is also what stops one flaky route creating a false page.
dated change
Cloud Monitoring starts charging for alerting on September 1, 2027
Google's pricing page carries a dated line most GCP monitoring articles have not caught up with yet. From September 1, 2027, Cloud Monitoring begins charging for alerting policies: $0.35 per month for each metric reference in an alerting policy, plus $0.50 per 1,000,000 points returned by the query of a metric alerting policy condition. Google says customers will receive notices 90 days and 30 days before the pricing is enabled, and that anyone on an active commit contract with discounts as of May 31, 2026 is exempt until that contract expires, renews, is extended or is replaced.
Here is the part to get right, because it is easy to get wrong and several summaries already have. Metric alerting policies that use Billing, Quota or Uptime metrics are not charged. Your uptime check alerts specifically stay free after September 2027. What starts costing money is everything else you alert on: latency, error rates, queue depth, custom application metrics, each metric reference in each policy.
The practical consequence is a shape change rather than a bill shock. Teams with a handful of policies will barely notice $0.35 a month per reference. Teams that generated hundreds of policies from Terraform modules, one per service per environment, will find that a number nobody has ever counted is suddenly a line item. Counting them is a twenty minute job today and a budget conversation in 2027.
the mechanics
What a Google Cloud uptime check actually does
Public uptime checks query an application that responds to HTTP, HTTPS or TCP requests, and can also send ICMP pings, up to three per check. That covers most of what a buyer means by uptime monitoring. The gaps are worth knowing before you commit, and they differ between the public and private check types in ways the console does not surface.
Public uptime checks
- HTTP, HTTPS and TCP, plus up to 3 ICMP pings per check.
- Period of 60s, 300s, 600s or 900s. Timeout 1 to 60 seconds.
- Issued from checkers in the USA (Oregon, Iowa, Virginia), Europe, South America and Asia Pacific.
- Selecting Global issues the check from every region, and bills every region.
- Works against any publicly reachable endpoint, inside or outside Google Cloud.
- Capped at 100 uptime checks per metrics scope.
Private uptime checks
- Reach HTTP or TCP endpoints inside your VPC, honoring IAM and VPC Service Controls.
- Require Service Directory: a namespace, a service and an endpoint per target, with private network access enabled.
- Need the Cloud Monitoring, Service Networking and Compute Engine APIs turned on.
- SSL certificate validation is disabled, regardless of how you configure it.
- Redirections are not supported.
- Results can take up to 5 minutes to appear.
That fourth bullet on the right is the one to flag to whoever owns your certificates. If your internal services are monitored by private uptime checks, Cloud Monitoring is not watching their certificates at all, and a certificate that expires behind the VPC boundary will take a service down with no prior warning from the monitoring stack. Our SSL certificate monitoring runs on every plan including Starter, on public endpoints, and warns before expiry rather than after.
pick one
Which GCP uptime monitoring setup fits you
Under 7 URLs
Use Cloud Monitoring and pay nothing
A handful of endpoints at a 60 second period sits inside the free million with room to spare. Set three USA checkers rather than Global, attach a notification channel to the alerting policy, and you are done. Do not buy a second tool for this. We would rather tell you that than sell you a plan you do not need yet.
8 to 100 URLs
The meter overtakes a flat plan fast
This is the band where the arithmetic flips hard. Twenty URLs at 60 seconds is $488.99 a month on Google Cloud and $19 with us. Either drop the period to five minutes and accept the resolution loss, or move the checks to a flat plan and keep the minute.
Signed 99.95% or tighter
You need faster than 60 seconds
A 60 second period burns 5% of a 99.95% monthly budget on one missed check, and Google sells nothing faster. If you have to produce evidence a customer will accept, you need both a shorter interval and a measurement taken outside the platform being measured. SLA monitoring software compares the options.
straight answer
What AlertPing does not do for Google Cloud
We are an external checker, not an observability platform, and there is a long list of things Cloud Monitoring does that we do not attempt. We cannot reach inside your VPC, so there is no equivalent of a private uptime check. We collect no metrics from your instances, so no CPU, memory or disk, and no agent. We do not do APM, traces, logs, real user monitoring or scripted browser journeys, which means Google's synthetic monitors do a job we do not do at all.
We also cannot correlate a failed check against the stack trace behind it, because your logs live in Cloud Logging and we never see them. For a lot of teams the right answer is both: Cloud Monitoring for everything happening inside the platform, and an independent check outside it for the availability record and the page at 3am. What we do is check sites, APIs and ports, confirm every failure from three regions before alerting, and send SMS, email, Slack and webhook alerts with no credit packs, on flat plans from $19 a month, with 30 second checks from $59.
questions people ask
GCP uptime check questions
How much does a GCP uptime check cost?
Google charges $0.30 per 1,000 uptime check executions, with 1 million executions free per Google Cloud project each month. Because each region counts as its own execution, one URL at the default 60 second period across three USA checkers is 131,499 executions a month, or $39.45 once the free allowance is used up.
Are GCP uptime checks free?
They are free up to 1 million executions per project per month, which is genuinely free for small deployments. In practice that covers 7.6 URLs at a 60 second period from three USA checkers, or 3.8 URLs if the check is set to Global. At a 15 minute period it stretches to 114 URLs. Past those points you pay $0.30 per 1,000.
What is the minimum interval for a GCP uptime check?
60 seconds. The uptime check API accepts only four period values: 60s, 300s, 600s and 900s, and 60s is the default. There is no sub minute option at any price or tier. If you need a 30 second or 15 second check to evidence a tight availability target, Cloud Monitoring cannot produce it.
Does GCP have uptime monitoring?
Yes. Cloud Monitoring uptime checks are Google Cloud's built in uptime monitoring. They query HTTP, HTTPS or TCP endpoints on a schedule from checkers in several regions, can send ICMP pings, and feed an alerting policy. They work against any publicly reachable endpoint, including services hosted outside Google Cloud.
How do I create an uptime check in GCP?
In the Cloud Console open Monitoring, go to Uptime checks and select Create uptime check. Choose the protocol, enter the hostname and path, pick the period and the checker regions, then attach a notification channel to the alerting policy it creates. Without a notification channel the check records results but pages nobody.
How many uptime checks can I have in GCP?
100 per metrics scope. A public uptime check also accepts a maximum of 3 ICMP pings. If you need more than 100 endpoints you have to split across metrics scopes, which splits your dashboards and alerting policies too. At 100 checks running every 60 seconds from three USA checkers, the monthly bill is $3,644.97.
What is a private uptime check in GCP?
A check that reaches HTTP or TCP endpoints inside your VPC while enforcing IAM and VPC Service Controls. It requires Service Directory with private network access enabled, plus the Service Networking and Compute Engine APIs. Two limits catch people: SSL certificate validation is disabled regardless of configuration, and redirections are not supported.
Does Google Cloud charge for alerting?
Not yet. From September 1, 2027 Cloud Monitoring charges $0.35 a month per metric reference in an alerting policy, plus $0.50 per 1,000,000 points returned by a metric condition query. Alerting policies using Billing, Quota or Uptime metrics are not charged, so uptime check alerts stay free. Commit contracts active on May 31, 2026 are exempt until renewal.
Can GCP monitor a website hosted outside Google Cloud?
Yes. A public uptime check works against any HTTP, HTTPS or TCP endpoint reachable from the public internet, including a site on another cloud or a third party API you depend on. You still need a Google Cloud project to hold the check, and you are still billed per execution per region for it.
What is the difference between a GCP uptime check and a load balancer health check?
A load balancer health check exists to make routing decisions, so when a backend fails it quietly removes it from the pool and tells nobody. An uptime check exists to record and alert. A service can look perfectly healthy to the load balancer, because failover worked, while a whole region of customers sees errors.
Watch your Google Cloud endpoints from outside Google Cloud
AlertPing checks your Cloud Run services, load balancers, App Engine apps and API endpoints as often as every 30 seconds, confirms every failure from three regions before it counts as downtime, and sends SMS, email, Slack and webhook alerts on every plan with no credit packs. Flat pricing from $19 a month, with 30-second checks from $59.
keep reading