What "Grok status" actually measures
Our automated monitoring probes grok.x.ai from a US-based data center using direct HTTPS requests
every 60 seconds. Each check captures four data points: the HTTP status code returned by Grok's servers,
the time-to-first-byte latency in milliseconds, whether DNS resolution succeeded, and whether the connection
completed within our timeout window. A single failed probe is not enough to mark Grok as down — we
require two consecutive failures before declaring an outage, which prevents transient network
blips from triggering false-positive alerts. This methodology is documented in detail on our
methodology page.
Why "Grok down for everyone or just me" is the right question to ask
When Grok appears unreachable from your device, there are three distinct possibilities, each with
a different fix:
- Grok is down for everyone — confirmed by this page showing red status plus
elevated user reports. No client-side fix will help; the site operator has to resolve it. In this case
our live outages feed will list Grok alongside any other affected sites.
- Grok is down for your region or ISP — typical pattern is this page showing
green (our US monitoring server can reach it) while users on certain ISPs or in certain countries can't.
Try a different network (cellular instead of Wi-Fi, or vice versa), switch your DNS resolver to
1.1.1.1 or 8.8.8.8, or test from a VPN to a different region.
- Grok is down only for you — local issue: corrupted DNS cache, browser
extension interfering, an outdated TLS certificate cached, or a corporate firewall. Fixes: flush DNS
cache, try incognito mode, try a different browser. If others on the same network can reach
Grok, the problem is on your machine.
Grok response time and what it tells you
The current response time for Grok is 63ms. For context:
a healthy Grok response is typically under 500ms; values rising into the 1500-3000ms range
usually indicate the site is under heavy load and may transition to a "Possible Problems" or
"Problems Detected" status within minutes. If you're seeing rising response times here, that's often a
leading indicator that an outage is brewing — refresh the page in a few minutes for the next data point.
The 24-hour response-time chart further down this page shows the full latency history.
For a deeper explanation of how response time fits into web performance, see our
TTFB explainer.
Grok in the Artificial Intelligence category
Grok is one of the artificial intelligence services in our 2,700+ monitored catalog.
artificial intelligence services share characteristic outage patterns — see the
Artificial Intelligence category page for an aggregated
"all systems operational" snapshot across every artificial intelligence site we monitor.
When a major artificial intelligence platform has an outage, related services on similar infrastructure
often see correlated issues, so checking the category view alongside this page gives a fuller picture.
How to report a Grok outage
If you're experiencing problems with Grok that aren't yet reflected above, scroll to the
User Reports section and submit your issue — it takes 5 seconds, no signup required.
Reports are anonymized (your IP is hashed with a daily rotating salt) and aggregate against a 7-day baseline.
When reports for Grok spike 3× above normal, the status flips to "Possible Problems";
a 5× spike or a confirmed HTTP failure flips it to "Problems Detected". Your report directly
contributes to the signal that other users see seconds later.
Embedding Grok status on your own page
A live SVG status badge for Grok is available at
https://isitdownchecker.com/api/badge/grok.x.ai.svg — drop it into your README, a status page,
or any blog post. The badge color updates every 60 seconds and links back here. Full embed instructions
and copy-paste snippets for Markdown, HTML, and BBCode are on our
embed page.
Common error patterns when Grok fails
Real outages on Grok don't always look the same to users. Common patterns we see in crowd reports:
- HTTP 5xx errors (502 Bad Gateway, 503 Service Unavailable) — server-side; see our
HTTP 502 explainer and HTTP 503 explainer.
- DNS resolution failures — site won't load, browser shows "server not found"; see our
DNS resolution explainer.
- TLS / certificate errors — site loads but browser blocks it; see our
SSL/TLS errors explainer.
- Loading hangs without an error — usually elevated TTFB or stalled connections; see our
TTFB explainer.
- Specific subsystem failures — login works but search broken, video plays but chat
fails, etc. — these are the cases where crowd reports surface the issue before HTTP probes do.