Your website doesn’t run in isolation. It loads JavaScript from CDNs, fonts from Google, payment forms from Stripe, analytics from third-party vendors. When any of these fail, your users see a broken page — even though your server is perfectly fine.
Website dependency monitoring is the practice of tracking every external resource your site loads and alerting you when any of them fails. Unlike traditional uptime monitors that only check if your server responds, dependency monitoring checks what your users actually see.
Key takeaway: 93% of websites load at least one third-party resource (HTTP Archive, 2024). A single failed dependency can break your site while your uptime monitor reports 100% uptime. Dependency monitoring closes this gap.
Why traditional uptime monitoring falls short
Traditional uptime monitors work by sending an HTTP request to your server and checking the status code. If the server responds with 200 OK, it’s marked as “up.” This approach was designed for a simpler web — when a website was a single server returning HTML.
In 2026, a typical web page is an assembly of resources from dozens of external domains. According to the HTTP Archive 2024 Web Almanac, the median website loads resources from 27 unique third-party domains. The top 1,000 websites request content from 66 third-party sources on average.
Your server can respond perfectly while any of these dependencies fail:
| What fails | What users see | What your monitor reports |
|---|---|---|
| CDN serving CSS/JS | Unstyled or non-functional page | 200 OK |
| Google Fonts | Invisible text for 3+ seconds | 200 OK |
| Payment API (Stripe, PayPal) | Checkout doesn’t work | 200 OK |
| Analytics script | Broken tracking, missing data | 200 OK |
| Authentication provider | Users can’t log in | 200 OK |
Every one of these is a real incident that happened to real companies. The Fastly CDN outage in June 2021 took down Amazon, Reddit, Twitch, GitHub, and thousands of other sites for 49 minutes (Fastly Post-Mortem). Every affected site’s server was healthy. Every traditional uptime monitor reported “all good.”
How dependency monitoring works
Dependency monitoring adds a critical step to the monitoring process. Instead of just checking the status code, it:
- Fetches your page — sends a request like a real browser
- Parses the HTML — extracts every
<script>,<link>,<img>, and other resource references - Checks each resource — verifies that CDNs, fonts, scripts, and stylesheets are loading correctly
- Alerts on failures — notifies you when any dependency is down, slow, or returning errors
This gives you visibility into the full chain of resources that your page depends on — not just your server.
A typical website loads between 30 and 80 external resources. E-commerce sites often exceed 100, with payment processors, product image CDNs, review widgets, analytics, and marketing scripts all loading on every page.
Real-world incidents that dependency monitoring catches
CDN outages
Cloudflare handles 20% of all web traffic globally. When Cloudflare had a P0 incident in June 2022, thousands of websites went down for 90 minutes — including sites that didn’t even know they depended on Cloudflare through nested third-party scripts.
The AWS S3 outage in February 2017, caused by a single mistyped command, cost S&P 500 companies an estimated $150 million. 54 of the top 100 internet retailers saw website performance degrade by 20% or more.
Third-party script failures
Third-party JavaScript accounts for 30.5% of all requests on the median web page (HTTP Archive, 2024). A blocking script that fails to load can prevent the entire page from rendering.
Common third-party failure modes:
- Slow loading — a script that takes 10 seconds adds 10 seconds to your page load time
- DNS failures — the third-party domain’s DNS stops resolving
- Certificate expiry — the CDN’s SSL certificate expires, causing mixed content errors
- Region-specific failures — a CDN works in the US but fails in Asia
Font loading issues
When Google Fonts or a custom font CDN is slow or unavailable, browsers either show invisible text (FOIT) or fall back to system fonts (FOUT). Your page is technically “up” but the user experience is broken — especially for brands where typography is part of the design.
What to look for in a dependency monitoring tool
Not all monitoring approaches are equal. Here’s what actually matters:
Automatic resource discovery — the tool should parse your HTML and find dependencies automatically, not require you to list every URL manually. Your dependencies change every time you add a widget, update a library, or your CMS loads a new plugin.
Multi-region checking — a CDN might work in Europe but fail in Asia. Checking from a single region gives you a false sense of security. Look for monitoring from at least 5-6 global regions.
Resource-level alerting — you need to know which specific resource failed, not just “something is wrong.” Did Stripe.js fail to load? Is your image CDN slow in Singapore? Which font file is timing out?
Correlation with main site — dependency checks should run alongside your main uptime check, not as separate monitors you have to configure individually. When a dependency fails, you want to see it in the context of your site’s overall health.
Dependency monitoring vs synthetic monitoring
Synthetic monitoring (tools like Datadog Synthetics, Pingdom Transaction Monitor) runs scripted user journeys — “click this button, fill this form, check this element.” It’s powerful but complex:
| Aspect | Dependency monitoring | Synthetic monitoring |
|---|---|---|
| Setup time | Minutes (automatic discovery) | Hours (write scripts per flow) |
| Maintenance | Zero (auto-discovers new deps) | High (scripts break on UI changes) |
| Coverage | All external resources | Only scripted paths |
| Cost | Low | High (per-step pricing) |
| Best for | External resource failures | User flow validation |
Dependency monitoring and synthetic monitoring solve different problems. Dependency monitoring answers “are all the pieces of my page loading?” Synthetic monitoring answers “can a user complete this specific task?” For most websites, dependency monitoring covers 90% of the failure modes at a fraction of the cost and complexity.
Getting started with dependency monitoring
upsonar monitors your website and all its external dependencies from 9 global regions. It automatically discovers every CDN, script, font, stylesheet, and API your pages load — no manual configuration required.
You can try it right now without creating an account:
- Dependency scanner — see all external resources on any website
- Website availability checker — check your site from multiple regions
Or create a free account — the free plan includes dependency monitoring for 3 websites.
Related: Beyond Soft 404s: 5 Ways Your Website Can Return 200 OK But Still Be Broken — more failure modes that status code monitoring misses.