HTTP Check
probe.http sends an HTTP or HTTPS request to each target of the step and asserts the response. Use it to verify a management UI, REST API, or health endpoint answers correctly — a status code in the expected set, optionally a body containing expected text — without any device credentials.
Using It
Targets come from the step's target picker: each selected role resolves to its devices' management addresses, and explicit IP selectors add literal addresses (an IPv6 address is put in brackets in the URL; one already in brackets is used as is). The request URL is built from Scheme (HTTP by default), Port (defaulting to 80 or 443 by scheme), and Path (default /; a missing leading slash is added). Method is GET, HEAD, or POST — keep in mind a HEAD response has no body for Body Contains to match.
Expected Status lists acceptable status codes as comma-separated values and ranges, for example 200-299,301; the default 200-399 accepts success and redirect responses. Body Contains, when set, additionally requires the response body to contain that exact text. Follow redirects (on by default) chases redirects and asserts the final response; switch it off to assert the redirect response itself. Verify TLS certificate (on by default) applies to HTTPS: it checks the certificate against the public certificate authorities plus the organization's trusted CAs (Settings, Trusted CAs), and can be switched off for self-signed device certificates. Timeout (sec) bounds each request attempt (default 10; steps saved with its old name, timeout_sec, keep working), and Attempts (default 1, up to 10) retries failing targets with a half-second pause between tries. Max Parallel Devices sets how many targets are requested at the same time (default 10, up to 32); 1 requests them one after another. Evidence and metrics always follow target order.
When the flow runs under an egress policy, every request — each redirect hop included — is checked against it, and a denied destination raises the policy's reason as the step error rather than counting as a failed check; the requests still running are then stopped.
Output
The summary reads http <method> <path>: N/M targets OK and is available to later steps as steps.NODE_ID.summary. Metrics record one <device name>_ms request latency per successful target. Each target adds one transport_meta evidence record named http_<device name> capturing the full URL, the method, the response status and latency, and the error text when the check failed.
Later steps read each target's result from steps.NODE_ID.output.targets['<device name>']: ok, error (empty when it passed) and latency_ms, plus the facts its evidence records (the url, method and response status). A target that was never checked, for want of an address, still appears with ok false and its error. Inventory names are unique per site only, and an IP target is named by its address: a name two targets share is keyed as name (device id), so every result is kept.
When It Fails
The step fails before any request when Expected Status cannot be parsed, or when no target devices or IPs are configured. A target fails when its address does not make a valid URL, when the request errors (refused, TLS failure, or timeout — the timeout applies per attempt, not to the step as a whole), when the response status is outside the expected set, or when the body does not contain the Body Contains text; the other targets are still requested. One failing target fails the whole step, with per-target reasons listed in the step error. Evidence is kept for every target that was actually requested; a device without a management address fails the step but leaves no evidence record.