When a website will not load, people reach for one of three kinds of tools: a crowd-sourced outage tracker such as Downdetector, an "is it down" checker that tests a URL on demand, or an uptime monitor that checks continuously and alerts you. They look similar, since all of them tell you whether something is "down", but they work in very different ways and fail in different ways. This page explains the difference so you know how much to trust each answer. SiteDownCheck provides the second and third kinds, and we will say where each approach is weak, including ours.
Approach 1: crowd-sourced reports (Downdetector and similar)
A crowd-sourced tracker does not necessarily test anything itself. It counts reports from people. As described on Downdetector's Wikipedia entry, it relies on user outage reports collected from sources including its own pages for each service and social media, and shows a map of where reports come from. The same article says it tracks well over ten thousand services across dozens of countries. We could not read Downdetector's own methodology pages in this session, so we cannot describe its exact detection rules, thresholds or any additional signals it may use. See their site for specifics.
What this model is good at:
- Big consumer services. When millions of people use a service, a spike of reports is a fast, human signal that something is wrong, often before the company posts anything.
- Regional patterns. A map of where reports come from can show that a problem is concentrated in one area.
- Problems no probe would find. A login failure for some users or a broken mobile app can show up in reports even when the homepage returns 200.
Where it is weak:
- Small sites have too few users to generate a signal. If your own site has a hundred visitors a day, nobody will file a report.
- Reports reflect perception, not measurement. A user's own Wi-Fi, an ISP issue or a service they misattribute can all produce a report. A news story can cause a spike of reports with no real fault.
- It cannot tell you why. A report count does not say whether DNS, TLS, the origin server or a CDN failed.
- It cannot alert you about your own site unless that site is one it tracks.
Approach 2: active probe, "is it down" checkers
An on-demand checker, such as our own HTTP status checker, makes a real request from its servers to the URL you enter and shows what happened: DNS lookup, connection, TLS handshake, HTTP status, response time and redirects. It is a measurement, not an opinion.
What it is good at:
- Answering "is it me or them?" in seconds. If the probe gets a 200 and you do not, the problem is probably your network, device or DNS cache. See website down for everyone or just me.
- Diagnosis. Our root-cause finder reports which layer failed: DNS, TCP, TLS, HTTP or content.
- Any site. It works for a site with no audience, because it does not depend on user reports.
Where it is weak:
- One vantage point. A probe from one location sees what that location sees. Our checks run from our own servers at a single default location, so a regional outage or a block aimed at our IP range can mislead us. Multi-region checks are planned for our Pro plan, which has no launch date. Down from where explains how to interpret results from one probe.
- A snapshot, not history. It tells you about the moment you pressed the button, not about last night.
- Probes can be treated differently from real users. Bot protection, rate limits and geo rules may block a data center IP while real visitors pass. A 403 or 429 to a probe is not always an outage.
Approach 3: continuous uptime monitoring
A monitor automates approach 2 on a schedule and tells you. Our free plan checks 3 monitors every 5 minutes or slower, re-checks a failure before alerting, and sends email, Slack, Discord or webhook alerts. It counts connection failures, timeouts, certificate errors and HTTP 5xx, 404 and 410 as down; a 401, 403 or 429 counts as up with restricted access. You can also require a keyword or a JSON value.
What it is good at: knowing before your customers, a history of incidents, SSL and domain expiry warnings, and heartbeat checks for cron jobs.
Where it is weak: the interval (up to five minutes of blindness on our free plan), the single location, and the fact that it only tests what you told it to test. A monitor of your homepage will not notice that checkout is broken.
Side-by-side
| Crowd-sourced reports | On-demand checker | Uptime monitor | |
|---|---|---|---|
| Source of truth | What users report | Active request from a server | Scheduled active requests |
| Works for tiny sites | No, needs many users | Yes | Yes |
| Tells you why | No | Often (by layer) | Often (error type) |
| Alerts you | Not for your own site unless tracked | No | Yes |
| History | Report trends | None | Incident log and uptime |
| Detects partial user-side faults | Sometimes | No | Only if you test that flow |
| Main risk | False reports, amplification | Single vantage point | Interval gaps, single location |
How to combine them
- Something looks wrong. Check a crowd-sourced tracker for a big third-party service, and an on-demand checker for your own or a small site.
- If the probe says up and you say down, test your own DNS and network: flush DNS, try mobile data, try another resolver. Our DNS checker और port checker help.
- If both say down, find the layer with the root-cause finder and follow the matching guide, for instance why is my website down.
- To stop finding out from users, add a monitor with alerts, plus a status page.
- For important URLs, use two independent monitors on different networks. How to reduce false alerts explains the logic.
Do not trust any of them blindly
None of these is an authority on its own. A crowd can be wrong, a probe can be blocked, and a monitor can be misconfigured. Treat each result as evidence. If three independent sources agree, you can act; if they disagree, the disagreement is itself the clue, and usually points to location, DNS or caching.
अगले कदम
- Try the HTTP status checker on any URL.
- Read website down for everyone or just me.
- Run the root-cause finder on a failing site.
- Start a free monitor so you hear first.
- Run a website health check.
सामान्य प्रश्न
Is Downdetector accurate?
It reflects what users report. That is a useful early signal for large services but not a measurement of any particular server, and it is less useful for small sites with few users.
Why does an "is it down" checker say up when my browser fails?
Because the probe runs from a different network. The problem is often local: your DNS cache, ISP, VPN, firewall or browser. Test from another network to confirm.
Can a checker be wrong?
Yes. One location can be blocked, rate limited or routed differently from your visitors. Multiple locations and repeated checks reduce this.
Does SiteDownCheck use crowd-sourced reports?
No. Our results come from active checks we run ourselves, from a single default location on the free plan.
Try SiteDownCheck free
3 monitors, 5-minute checks, email and webhook alerts and a public status page. No credit card.
मुफ़्त मॉनिटरिंग शुरू करेंOpinions here are our own; verify vendor details yourself. See our संपादकीय नीति और अस्वीकरण.