Validators allow you to run checks on the probe request output (if any). For example, you can specify if you expect the probe output to match a certain regex or return a certain status code (for HTTP). You can configure more than one validators and all validators should succeed for the probe to be marked as success.

probe {
    name: "google_homepage"
    type: HTTP
    targets {
        host_names: ""
    interval_msec: 10000    # Probe every 10s

    # This validator should succeed.
    validator {
        name: "status_code_2xx"
        http_validator {
            success_status_codes: "200-299"

    # This validator will fail, notice missing 'o' in our regex.
    validator {
        name: "gogle_re"
        regex: "gogle"

(Full listing:

To make the debugging easier, validation failures are logged and exported as an independent map counter – validation_failure, with validator key. For example, the above example will result in the following counters being exported after 5 runs:

total{probe="google_homepage",dst=""} 5
success{probe="google_homepage",dst=""} 0
validation_failure{validator="status_code_2xx",probe="google_homepage",dst=""} 0
validation_failure{validator="gogle_re",probe="google_homepage",dst=""} 5

Note that validator counter will not go up if probe fails for other reasons, for example web server timing out. That's why you typically don't want to alert only on validation failures. That said, in some cases, validation failures could be the only thing you're interested in, for example, if you're trying to make sure that a certain copyright is always present in your web pages or you want to catch data integrity issues in your network.

Let's take a look at the types of validators you can configure.

Regex Validator

Regex validator simply checks for a regex in the probe request output. It works for all probe types except for UDP and UDP_LISTENER - these probe types don't support any validators at the moment.

HTTP Validator

HTTP response validator works only for the HTTP probe type. You can currently use HTTP validator to define success and failure status codes (represented by success_status_codes and failure_stauts_codes in the config):

  • If failure_status_codes is defined and response status code falls within that range, validator is considered to have failed.
  • If success_status_codes is defined and response status code does not fall within that range, validator is considered to have failed.

We can possibly add more checks to HTTP validator in the future, for example to verify headers.

Data Integrity Validator

Data integrity validator is designed to catch the packet corruption issues in the network. We have a basic check that verifies that the probe output is made up purely of a pattern repeated many times over.