"Is it down for everyone or just me?" is the first question after any failed page load. The honest answer is that the pattern of where a site fails tells you almost everything, and a single probe from a single machine can't reveal that pattern. Three quick checks can.
Three possibilities, one decision tree ¶
When a site doesn't load, exactly one of three things is happening:
- Global outage, the provider is down for everyone worldwide.
- Regional outage, users in specific geographies can't reach it; others are fine.
- Local issue, only your machine, browser, or network is affected.
Each leaves a distinct fingerprint. The 60-second diagnostic below eliminates the possibilities in order of cost-to-test: the fastest, cheapest check first.
Step 1, Multi-method check (10 seconds) ¶
Open isitdown.io and probe the target using 4 independent probe methods in parallel. The result:
- All 4 probe methods fail, confirmed global outage. Nothing you do locally will fix it.
- Some regions fail, others succeed, regional outage. If you're in a failing region, a VPN exit in a healthy region usually restores access.
- All 4 probe methods succeed, the outage is local to you. Continue to step 2.
A multi-method check beats a single-region tool like ping or curl from your laptop because one machine can't distinguish "the site is down" from "you can't reach the site." Why that matters →
Step 2, Flush DNS (15 seconds) ¶
A surprising fraction of "down for me" reports are stale DNS cache entries on the user's machine. When a site migrates to new infrastructure, your OS may keep the old IP cached for minutes or hours.
- Windows, Command Prompt:
ipconfig /flushdns - macOS, Terminal:
sudo dscacheutil -flushcache; sudo killall -HUP mDNSResponder - Linux (systemd-resolved),
sudo resolvectl flush-caches - Chrome, visit
chrome://net-internals/#dnsand click "Clear host cache."
Retry the page. If it loads, you were looking at a cached IP.
Step 3, Browser and network reset (20 seconds) ¶
If DNS isn't it, the problem is one of:
- A bad extension intercepting the request (ad blockers, VPNs, security plugins).
- A corrupted service worker serving a cached error page.
- An ISP or corporate filter blocking the domain.
- A router-level DNS issue (some home routers cache aggressively).
Quick test matrix:
- Incognito / private window, bypasses extensions and most cached state.
- Different browser, isolates browser-specific issues.
- Mobile hotspot, if the site works on cellular but not home Wi-Fi, the issue is your router, ISP, or home DNS.
- Public DNS, temporarily switching to
1.1.1.1(Cloudflare) or8.8.8.8(Google) rules out your ISP's resolver.
When to wait vs. when to act ¶
| Pattern | Cause | What to do |
|---|---|---|
| All methods fail, provider confirms | Global outage | Wait. Subscribe to their status page for updates. |
| All methods fail, provider says "all good" | Monitoring blind spot | Wait 10-15 minutes. Provider dashboards often lag. |
| Some methods fail, others pass | Request-type block / partial | The site is up for browsers; try a normal browser visit. |
| All methods pass, only you affected | Local | Run the three steps above. |
FAQ ¶
How often do "down for me" issues turn out to be local?
In direct-probe data, a meaningful fraction of "is X down" queries from the US and Europe resolve to local or regional issues rather than global outages. The exact ratio varies by provider and by time of day.
Do third-party "is it down" sites crowd-source reports?
Most do, they aggregate social-media mentions and user self-reports. That's fast but biased: a trending topic produces spikes even when the provider is healthy. isitdown.io uses direct probes from our monitoring infrastructure, so there are no crowd-reported false positives.
What counts as a "confirmed outage" on isitdown.io?
In our classification, a confirmed outage is when all four probe methods fail within the same five-minute window. Anything less is partial or local until the pattern holds across more than one window.