Edge · Use cases

Seven ways storefronts get hurt, and how Edge sees each one.

The through-line for all seven: Edge reads your storefront traffic server-side, from the Salesforce B2C Commerce eCDN log itself, with no JavaScript tag on your pages. That is why it sees the clients a page tag never can.

Bot impersonation

The scraper wearing Googlebot’s badge

You allowlist Googlebot so your SEO is never blocked. Price scrapers know that, so they set their User-Agent to Googlebot and walk straight through the one identity you can never turn away. Your tag-based bot tool never sees them, because a scraper does not run JavaScript.

How Edge sees it

Edge reads your traffic from the server side, from the eCDN log. It shows you the requests that claim to be Googlebot but that Cloudflare could not verify, grouped by the network they came from, and gives you a ready-to-apply challenge rule scoped to that network and that claimed identity. Genuine Googlebot, on Google’s own network, is never touched. Nothing goes live until you approve it.

Proof you can verify

Real crawlers verify on their own networks. Impersonators run on rented cloud servers. You can confirm the split in your own eCDN logs, and confirm the same requests are missing from your tag vendor’s console, because the client never ran the tag.

See who is impersonating your crawlers

Payment and login abuse

Card testing you cannot see, from ten thousand addresses

A card-testing operation validates stolen cards against your checkout. A credential-stuffing run hammers your login. Both spread thin across thousands of addresses, so no single one looks abnormal, and both send raw requests that never render a page. A page tag never fires, so a client-side tool shows you nothing.

How Edge sees it

When several independent signals point at the same endpoint at once, Edge raises one incident instead of a cascade of lookalike alerts. From it you drop into the exact controller under pressure, see the addresses behind it, and take a paste-ready blocklist. You act on the real attack, human-approved, and leave a record of what you did.

Proof you can verify

The connections a client abandons before your page loads, and the raw POST submissions to your endpoints, are in your eCDN logs and, by construction, in no page-tag tool. Confidence here is independent detectors agreeing, not one vendor’s opaque score, and you can click each one to see the evidence.

Turn scattered noise into one clear incident

Retail peak events

Survive Black Friday without silencing your fraud detectors

A drop or flash sale is coming, and the surge will look exactly like the abuse your volume detectors are built to catch. Pause alerts globally and you also blind yourself to the card testers and scalpers who deliberately hide behind a sale. Leave them on and on-call gets paged all night.

How Edge sees it

Declare the window in advance. Edge holds the volume alarms on the expected surge, for the source and dates you choose, while your fraud detectors stay live. The detectors an attacker most wants muted cannot be suppressed at all. Afterward, a per-event report shows how much of the surge was automated and how much never reached your tag analytics.

Proof you can verify

The list of what can and cannot be held is visible in the product: try to suppress a card-testing or credential detector and you cannot. The automated-share figure comes from your server-side logs, so it counts the non-human traffic a page tag misses during your busiest hours.

Plan your next drop without the 3am pages

Scraping and origin cost

The scraper lifting your prices and inflating your bill

A cloud-hosted scraper pulls your product and search pages at machine speed and bypasses cache to always read fresh prices. It lifts your catalogue, undercuts you, and inflates the origin traffic you pay for. It never runs your tag, so a tag-based tool cannot rank it or name its network.

How Edge sees it

Edge ranks the networks hitting you by the signals that separate a scraper from a shopper: how much they bypass cache, how small and machine-like their requests are, and whether they come from a hosting network rather than a home connection. One click builds the case against the worst offender and hands you a blocklist ready for your bot rules.

Proof you can verify

The ranking counts every request from the eCDN log, including the raw clients a tag never sees, and it carries Cloudflare’s own cache decision and firewall action per network, which a client-side tool cannot know for a client that never loaded it.

Find the network eating your catalogue

Analytics integrity

Your funnel numbers are computed against the wrong population

A large slice of your traffic is API clients, declared crawlers and scripts with no browser at all. None of it fires your analytics tag, so your conversion rate, your bounce rate and your channel numbers quietly exclude a big, non-human chunk of what actually hit your store. The tool that produces those numbers cannot measure its own blind spot.

How Edge sees it

Edge puts a number on it. It splits your traffic into the requests that could never fire a tag and the browser-shaped traffic that could, and flags the browser-shaped but automated slice that is polluting your reports. Then it names the AI agents and crawlers behind it.

Proof you can verify

The “never fires a tag” figure is exactly the traffic missing from your Adobe or GA4 reports for the same window. Pull both and see the gap.

Find out how clean your funnel data really is

Early warning

Someone is casing your store before they attack it

Before a targeted attack, someone fingerprints your storefront: probing for exposed secrets, admin panels, config files, and filenames built from your own brand name. It is low volume and easy to miss, and it is the tell that you specifically are on a list. These probes hit dead-end URLs that render no page, so a JS tag never runs and is blind to reconnaissance entirely.

How Edge sees it

Edge captures the failed probes a normal report buries, sorts them into named scanner kits, and ranks them by what was tried and who tried it. A probe path that carries your own brand name is treated as critical no matter how quiet it is, because that is a kit built for you, not a spray hitting everyone.

Proof you can verify

The probes are in your eCDN logs and carried no tag execution, both checkable in your own data. Sorting them into scanner families and flagging your brand name is work a raw log tool does not do.

See who is casing your store

Uptime and outages

The two failures your dashboards hide until customers call

Two quiet outages slip past a normal dashboard. A TLS certificate on a storefront hostname lapses, and every browser throws a security warning, so checkout stops. Or your log stream dies, and every dashboard keeps showing yesterday’s data as if all is well, so your detection has gone dark with no red light. In both cases you find out from shoppers, after the outage starts.

How Edge sees it

Edge is an independent watchdog. It checks the certificate your storefront is actually serving, with a live probe where the platform allows one and public certificate records as the cross-check, and counts down to expiry, warning you with days of lead time. And it watches your log stream itself, so a dead feed raises an alert instead of a stale but green dashboard.

Proof you can verify

Certificate monitoring cross-checks the publicly logged certificate, so you can look up any storefront hostname on a public transparency log and confirm what Edge is watching, regardless of who manages the certificate inside the platform. The tag vendors Edge is sold against do not monitor certificates at all.

Get warned before your customers are

See all seven against your own traffic.

Connect your Logpush and see your first dashboard in minutes. No tag, no code change, no sales call. Nothing happens to your storefront that you didn't decide.

Start 30-day free trial