Check pricing
Overview
Checkly supports different kinds of checks which are billed slightly differently:
Check type | Base cost | Parallel scheduling | Retries |
---|---|---|---|
Browser | $5 per 1k runs | Each location is a run | Each retry is a run |
API | $2 per 10k runs | Each location is a run | Each retry is a run |
Multistep | $2 per 10k requests | Locations x requests | Retries x requests |
Heartbeat | - | - | - |
When configuring your check frequency and scheduling strategy, the cost helper will estimate the monthly cost for the check.
API and Browser checks
Browser checks and API checks have a base price per check run. Always refer to our pricing page for the latest prices. The base cost is typically listed as $X per 1k or 10k check runs.
If you use features that multiply the number of check runs, such as parallel scheduling and retries, your cost will increase.
Parallel scheduling
When a check is running in parallel, it will run once on each selected location per execution. If you change a check from using the round-robin scheduling method to running in parallel, be aware that the cost will multiply by the number of locations you have selected.
Balance the number of locations the check is using to ensure you are quickly made aware of potential issues in critical locations without running an unnecessary amount of checks.
Retries
When a check is retried, this counts as a new check run. A flaky check can increase your check run costs above the expected. If you have problems with flaky checks, ask our community or our support for tips on how to improve check stability and reduce cost.
Multistep checks
Multistep check pricing is slightly different from browser and API checks. A Multistep check is billed based on the number of requests done per check run. Each request in a Multistep check run is billed as a single regular API check run, as they are performing the same basic operation.
GET
, POST
, PUT
and DELETE
operations towards the same endpoint. If you replace these 4 checks with a single Multistep check that runs 4 requests towards the same endpoint, checking each method, and the check run frequency is the same as before, your cost stays the same.A Multistep check with 0 requests is billed as if it has 1 request.
Parallel scheduling
When a Multistep check is run in parallel, the whole check (with any number of requests) is run from all included locations. That means that the cost is going to equal the number of locations the check is run from multiplied by the number of requests.
Retries
As a Multistep check is retried as a whole, the final cost of a retried Multistep check is based on the number of requests executed multiplied by the number of retries.
Heartbeat checks
A set number of heartbeat checks are included in the team and enterprise plans. Checkly does not charge for heartbeat pings.
If you have reached your maximum number of heartbeat checks and need more, contact our support.
Last updated on April 12, 2024. You can contribute to this documentation by editing this page on Github