Traffic Metrics

Fourteen-day traffic metrics for an account in vPanel.

Metrics (vPanel Dashboard → Preferences → Metrics) summarises traffic to all of your domains over the last 14 days, read directly from your web server's access logs. No tracking script, no JavaScript, no cookie — which means it counts things a client-side analytics tool cannot see, including bots, crawlers, API calls and requests that never rendered a page.

If you have no traffic yet, the page says so rather than showing an empty chart. Stats appear as soon as your sites start receiving requests.

The four headline numbers

• Requests — every logged request the panel could parse, across all your domains. This counts hits, not page views: one page load that pulls in twelve images and three stylesheets is sixteen requests.
• Unique visitors — distinct client IP addresses. A useful trend line, not a precise headcount: an office behind one NAT looks like one visitor, and a phone moving between networks looks like several.
• Bandwidth — total response bytes served.
• Success rate — the share of responses that were 2xx or 3xx, with the count of 4xx and 5xx responses beneath it.

Requests per day

The bar chart shows one bar per day for the last 14 days, ending today. Hover a bar for its exact date and request count. Days with no traffic show as empty rather than being skipped, so a gap is visible as a gap.

This is the fastest way to spot the shape of a problem: a cliff usually means DNS or an outage, a plateau at a suspiciously round number usually means something is being rate-limited, and a spike with a matching jump in 4xx usually means a bot.

Top pages

The ten most-requested paths with their hit counts. Paths are taken verbatim from the log, so query strings are kept and /page?a=1 and /page?a=2 count separately. Static assets are included here — if images dominate the list, that is real traffic, just not the traffic you were probably looking for. For a page-focused list with assets filtered out, use Visitors.

Responses

A breakdown of response codes into four buckets, each with its count and a bar showing its share of all responses:

• 2xx Success — the request worked.
• 3xx Redirect — normal for HTTP-to-HTTPS and for www handling. A very high share can mean a redirect loop or a redirect that should have been a rewrite.
• 4xx Client error — mostly 404s. A handful is normal. A steady stream means broken internal links, a missing favicon, or a bot probing for admin paths.
• 5xx Server error — your application failing. Any sustained 5xx share deserves attention; go to Logs and read the error log for the same period.

How the numbers are produced, and their limits

Metrics parses the Apache access logs belonging to your account under /var/log/apache2, in the combined log format. Two limits are worth knowing, because they explain surprising totals:

• Each log file is read up to 80,000 lines, counted from the start of the file. On a busy site the log is longer than that, and because the reading starts at the beginning it is the most recent requests that fall outside the window. So "Requests" is a floor rather than an exact total, and the newest days of the chart can look emptier than they really were. The page labels the figure last 14 days (capped) for that reason.
• Only logs still on disk are counted. Log rotation eventually removes old files, so on a server that rotates frequently the days that have rotated away drop off the chart entirely.

Requests that do not match the expected log line format are skipped, including some malformed or truncated entries.

Related pages

• Visitors — recent traffic in detail: referrers, IPs, user agents and individual requests.
• Logs — the raw access and error logs.
• Disk Usage — where your storage is going.