REST API Checker

Hit an API endpoint with GET or POST. See status code, latency, and whether it matches your expected response.

Request

Public endpoints only. No auth headers — use for health checks and open APIs.

Response

Awaiting request
Status code:
Latency:
Method:

When to use a REST API smoke test

Health endpoints (/health, /ready, /live, /status) exist so automation can ask “is the API process up?” without executing full user journeys. This checker sends one GET or POST from our server and compares the HTTP status code — optionally against your expected value.

Free tool, no signup. Honest scope: not API monitoring, not contract testing, not load testing.

Good use cases

  • Verify a deployed health URL after release from outside your VPC.
  • Confirm a public API gateway returns 200 when you suspect regional outage.
  • Compare GET vs POST behavior on a simple unauthenticated endpoint.
  • Demonstrate to stakeholders that external reachability works post-firewall change.

Worked example 1 — Kubernetes readiness

URL: https://api.example.com/ready · Expected: 200 · Result: PASS in 89 ms

Interpretation: load balancer path healthy from internet. Does not prove database migrations succeeded.

Worked example 2 — intentional 503 design

Some apps return 503 on /health when dependency down (fail-visible pattern).

Set expected status 503 during maintenance validation; default 2xx-only pass would misleadingly FAIL.

Worked example 3 — PASS but product broken

/health → 200. /v1/checkout → 500 for all users.

Lesson: add synthetic checks on critical business paths in proper monitoring; smoke test is floor not ceiling.

Not a substitute for

  • OAuth, API keys, or mTLS — no custom headers.
  • JSON schema or response body assertions.
  • Rate-limit and abuse testing.
  • Continuous multi-region synthetic monitoring.

When NOT to use

  • Internal microservices on private RFC1918 addresses.
  • GraphQL complex mutations requiring auth context.
  • Compliance audit trails — use CI with retained logs.

Designing health endpoints

Good /health returns 200 when critical dependencies are reachable — not always when every optional subsystem is up. Split /live (process up) and /ready (can serve traffic) for Kubernetes. Document which endpoint this checker hits in runbooks.

CI integration pattern

Run curl from CI after deploy; this page is for ad-hoc human checks from outside your VPC when CI green but customers report issues — different network path, same idea.

Deep dive: API gateways and edge auth

APIs behind API keys or JWT often return 401 to unauthenticated probes — expected, not “down.” Create dedicated unauthenticated health routes or use this tool only on public endpoints designed for anonymous checks.

GraphQL single endpoint may return 200 with errors array in body — this checker sees HTTP layer only. For GraphQL, extend monitoring in application tests that parse errors field.

Rate limiting can return 429 during aggressive CI; this tool sends one request — if 429, distinguish throttling from outage before paging.

Comparison: this checker vs curl vs Postman vs production monitoring

ToolBest forLimitation
This checkerQuick external GET/POST status from browserNo auth, one location, no body
curlScripted checks with headers from your laptopYour network path, not customer path
PostmanCollections, auth, assertionsManual or Newman CI setup required
Production monitoring24/7 multi-region with pagingCost and configuration time

Use this page when you want a shareable “external check failed at 15:04 UTC” screenshot for a war room without installing CLI tools on a support laptop. Escalate to CI and monitoring for permanent coverage.

HTTP status alone ignores response body contracts. An API returning 200 with {"status":"fail"} still breaks clients — follow with application tests. Latency shown here is one sample; cold starts and TLS negotiation can skew first request — run twice if borderline.

HTTP semantics methodology for health checks

REST smoke tests answer: “Did the server return the expected status code for this URL from outside?” Define pass criteria in runbooks before incidents. Common patterns: 200 = healthy, 503 = unhealthy but reachable (Kubernetes readiness), 401 = auth required (not necessarily down). Latency pass threshold example: fail if > 2000 ms for three consecutive external checks — this tool gives one sample; multiply manually or use monitoring for consecutive logic.

Common mistakes with API smoke tests

  • Checking /health while /v1/orders returns 500. Health endpoints often skip database depth.
  • Expecting 200 when API returns 401 by design. Set expected status explicitly in this tool.
  • Using POST here for JSON mutations. Empty POST body will fail validation on real APIs.
  • Declaring victory on 200 with error payload. Body may contain {"status":"degraded"} — HTTP layer only.
  • Ignoring 429 rate limits. One check throttled ≠ outage; retry after 60 seconds before escalating.
  • Testing from CI network only. Corporate egress may differ from customer path — external check still matters.

Worked example — latency regression with numbers

URL: https://api.example.com/v1/health · Expected: 200.

Baseline (last week): PASS in 95 ms, 102 ms, 88 ms (internal monitoring).

Today external check: PASS in 2,340 ms — still 200, but 22× slower than baseline.

Action: treat as degradation incident; open ticket before customers timeout at 3s client SDK limit. Full outage avoided; SLA latency SLO may still breach.

Honest limits

GET and POST without body, headers, or cookies. Single request from one datacenter. No GraphQL body parsing, no gRPC, no WebSocket upgrade checks. Cannot reach private RFC1918 hosts. Not a substitute for contract tests, schema validation, or load testing — a fast external sanity layer only.

Mapping check results to incident severity

External API check FAIL on /health during business hours often warrants SEV2 declaration if no workaround. PASS with latency 3× baseline may be SEV3 degradation — log in incident channel even without page. Document expected status codes per environment in runbook: staging may intentionally return 503 when dependencies are down for chaos tests — do not page production on-call with staging URL.

After firewall or WAF change tickets close, run this checker on three URLs: public health, one authenticated business path (expect 401 — confirms reachability), and admin path (expect 403). Pattern confirms policy applied without accidental block of legitimate monitor paths.

Versioning and deployment correlation

Tag each deploy with version header or /health body field when possible. When external check fails after deploy #8842, rollback decision is faster than binary search across six hours of commits. This tool cannot read response body — but HTTP status flip from 200 to 500 immediately after deploy timestamp strongly implicates release; note deploy ID in incident channel alongside checker result screenshot.

SLA-oriented API availability math

If contract defines API availability as successful 2xx on /health from external probe, one failed check does not breach monthly SLA — but 72 failed minutes in a 43,200-minute month = 99.83% achieved. Use continuous monitoring for vendor disputes; this checker validates “right now from outside” during active triage. Pair with Reliability Money Toolkit for customer-facing SLA math separate from vendor credits.

Webhook and callback endpoints

Payment and partner webhooks often use POST with signed payloads — this tool cannot validate signatures. For webhook reachability only, some teams expose unauthenticated /webhook-health returning 200. Document which endpoints are safe for anonymous external checks vs which must stay in CI with secrets.

Multi-step API diagnosis without this tool

OAuth login flows, file uploads, and idempotent PUT operations need CI integration tests — not browser smoke checks. Use this page at step zero when users report “API down” to confirm external TCP/TLS/HTTP path before diving into distributed traces. Saves 15–30 minutes when issue is WAF block vs application bug.

Documenting results for post-mortems

Paste checker output into incident timeline: URL, method, status code, latency ms, UTC timestamp, PASS/FAIL. Post-mortem readers six months later should reconstruct external perspective without re-running tools. Link to Post-Mortem Generator after resolution.

Related tools

Availability Checker · Speed Smoke Test · Downtime Cost · Free Tools.

Frequently Asked Questions

Can I send JSON body or auth headers with this API checker?

No. GET and POST only, empty POST body, no custom headers. Use Postman or CI for authenticated tests.

What is a REST API smoke test for?

Quick verification that a public health endpoint returns expected HTTP code from outside your network.

Why PASS with 200 but customers still broken?

/health may return 200 while business endpoints fail. Smoke tests prove one URL only.

Is this check run from multiple regions?

No. Single request from our server.

How do I set expected status code?

Optional field — if blank, 2xx passes. Some health endpoints return 503 when unhealthy by design.

Can I test localhost APIs?

No. URL must be reachable from the public internet.

Does POST send a request body?

POST sends an empty body. JSON APIs requiring payload will not work here.

What should I do after API check fails?

Verify DNS, SSL, availability; inspect logs; confirm WAF; run full integration tests in CI.