Backend Failure Detection
API & Backend Failure Detection
Your backend goes down. Your server returns a 503. Your CDN starts serving an error page. But your uptime tool? It sees a 200 OK. Sitewatch fetches pages and validates every asset, so backend failures that break the visitor experience get caught -- not ignored.
- Detects 5xx errors and service unavailability on your pages
- Catches redirect loops and pages silently drifting to wrong hosts
- Finds missing assets caused by backend or deployment failures
Backend failure detected
Confirmed via 2-of-3 retries
Affected pages
Recent activity
- /checkout -- 503 Service Unavailable6:00 AM
- /account -- redirect loop detected (3 hops)6:00 AM
- /products -- main.js returned 404 Not Found6:00 AM
- /about -- all 12 assets healthy6:00 AM
Why it matters
Backend failures that uptime tools pretend do not exist
Server error detection
When your backend returns a 5xx error, Sitewatch flags the page as UNAVAILABLE. That covers 503s, 502s, and network failures that your ping-based uptime tool never sees.
Redirect loop detection
A misconfigured backend traps visitors in an infinite redirect cycle. The browser gives up, but your server logs look normal. Sitewatch flags REDIRECT_LOOP before users get stuck.
Host drift detection
Your page silently redirects to an unexpected domain -- a CDN error page, a parked domain, a staging server. HOST_DRIFT catches it because Sitewatch follows redirects and checks where they land.
Broken asset detection
A backend deployment breaks asset references. JavaScript returns 404, images vanish, CSS files disappear. Sitewatch sends a HEAD request to every referenced asset and flags ASSET_MISSING failures.
Content type validation
When servers return wrong MIME types or serve non-HTML responses for pages that should be HTML, Sitewatch flags ASSET_MIME_MISMATCH and NON_HTML_PAGE incidents.
Structural change detection
SHA-256 fingerprinting of your page HTML detects when backend issues silently alter page output. You see when content changes unexpectedly, even if the status code stays 200.
6
Incident types detected
2-of-3
Retry confirmation
6 channels
Alert channels
11 detection rules
Every way a backend failure can break your site
Page-level failures
- UNAVAILABLE -- page returns 5xx error or network failure
- REDIRECT_LOOP -- page caught in circular redirects
- HOST_DRIFT -- page redirects to an unexpected domain
- NON_HTML_PAGE -- page serves wrong content type instead of HTML
Asset-level failures
- ASSET_MISSING -- JS, CSS, or image returns 404 or 410
- ASSET_MIME_MISMATCH -- asset served with incorrect MIME type
Simple setup
How Sitewatch catches backend failures
Add your pages
Enter any URL and Sitewatch fetches the page via HTTP, just like a real visitor would.
Parse and validate
The HTML is parsed to extract every script, stylesheet, and image reference. Each asset gets a HEAD request to check its HTTP status code and MIME type.
Confirm with retries
Issues are confirmed with 2-of-3 retry logic to eliminate false positives from transient glitches.
Get alerted via Slack or email
Confirmed incidents trigger alerts via Slack or email with the failure type, affected URL, HTTP status code, and which assets broke.
Stop backend failures from reaching your visitors
Free plan. No credit card. 2-minute setup.
The gap in your stack
Your uptime tool says "up." Your visitors say "broken."
| Feature | Uptime tools | Sitewatch |
|---|---|---|
| Server errors | Ping says 200 OK | Full page fetch with 5xx detection |
| Redirect issues | Not detected | Redirect loops and host drift caught |
| Asset validation | Not checked at all | Every JS, CSS, and image verified |
| Content type checks | Ignored | MIME type validated per asset |
| False positive prevention | Alert on first failure | 2-of-3 retry confirmation |
| Alert evidence | Status code only | Failure type, asset URLs, HTTP status |
Server errors
Redirect issues
Asset validation
Content type checks
False positive prevention
Alert evidence
FAQ
Frequently asked questions
Sitewatch monitors web pages, not arbitrary API endpoints. It fetches your pages via HTTP and validates that they load correctly -- catching 5xx errors, redirect loops, host drift, and broken assets. If your backend fails and it affects what visitors see, Sitewatch detects it.
Sitewatch fetches each page like a real visitor, then parses the HTML to find every script, stylesheet, and image. Each asset gets a HEAD request to verify its HTTP status and MIME type. If the page itself returns a 5xx error or redirects unexpectedly, that is flagged too.
Every detected issue goes through 2-of-3 retry confirmation. Sitewatch also uses a 30-minute per-incident alert cooldown and fingerprint-based deduplication -- one problem creates one incident, not a flood of alerts.
Sitewatch supports 6 alert channels: Slack, email, SMS, webhook, PagerDuty, and Opsgenie -- configured at the workspace level. Each alert includes the incident type, affected URL, and HTTP status codes so your team can triage fast.
Check intervals are configurable per site -- every 30 minutes on Free, every 5 minutes on Pro. On-demand checks are also available from the dashboard and via deploy hooks for immediate post-deploy verification.
Explore more