How-to
Reading a Path Detail Page
Priority badges, site-wide baseline overlay, 7-day history, HTTP methods, related alerts, and unique IPs per ASN — how to read what the page is telling you.
Jump to section
What the Path detail page shows
Every path prefix on the Paths list drills into a detail view with the full signal set for that endpoint. Use this page when you need to answer: "is this path being abused?".
Priority badge
Some paths carry a coloured Priority pill next to the title. These are SFCC controllers Edge treats as high-signal:
- Red — Checkout / Payment (COBilling, CheckoutServices, COSummary, Order, PaymentInstruments, Adyen) — card-testing, gift-card enumeration, payment fraud target
- Amber — Auth / Credential (Login, Account, Customer, ResetPassword) — credential stuffing, account takeover target
- Blue — Cart / Basket (Cart, BasketMgr) — grinch bots, inventory hoarding target
The badge is a cue to scrutinise traffic patterns on this path more carefully than you would a browsing endpoint.
Requests stat tile — "X× 7d avg"
The Requests tile's sublabel includes a 7-day-average comparison: "0.2% of total · 1.4× 7d avg (normal)".
- quiet (< 0.5×): traffic is well below typical
- normal (0.5–2×): within the expected range
- elevated (2–5×): unusually busy but not alarming
- spike (≥ 5×): well above typical — worth investigating
The 7-day baseline is independent of the currently-selected range — it always compares against the last 7 days' daily average.
Status-class timeseries — Site-wide overlay
The chart shows this path's requests broken down by HTTP status class (2xx, 4xx, 499, etc.), stacked. A dashed grey Site-wide line overlays the total site traffic shape.
- Path tracks the site baseline → this path is riding a general traffic pattern (legit campaign, time-of-day)
- Path diverges from the baseline → this path has independent behaviour; could be a targeted attack
Top ASNs table — Unique IPs column
The Top ASNs table shows, for each ASN hitting this path:
- Requests — volume from that ASN
- Unique IPs — distinct client IPs within that ASN captured in the path's top-N IP rollup
Interpreting the Unique IPs column:
- Red (≤ 2 IPs) — likely bot. One or two IPs carrying meaningful volume is a classic scripted-attack signature.
- Amber (3–5 IPs) — worth a look. Small egress pool could be a small bot cluster or a corporate proxy.
- White (≥ 6 IPs) — distributed traffic, likely legitimate humans.
- — (em-dash) — ASN didn't crack the top-N IP rollup for this path. Usually means low ASN volume — not a signal either way.
Note: this is "how many of this path's top IPs belong to this ASN" rather than a true distinct-count. For offender detection that's the signal that matters.
HTTP methods card
Colour-coded method breakdown:
- Green — read-only methods (GET, HEAD, OPTIONS). Safe and idempotent.
- Amber — mutating methods (POST, PUT, PATCH, DELETE). On prescribed-POST endpoints like COBilling-Submit, this bar should be 100% — any GET volume is probing.
- Red — (no method) — Logpush didn't supply a method for these requests. See HTTP methods and required fields for what to do about this.
High-signal patterns:
- Login-Submit showing significant GETs — someone is scraping the form / harvesting CSRF tokens
- Search endpoint flipping from 80% GET to 80% POST — unusual scraper behaviour or testing of a different attack vector
- Adyen-Notify getting GETs — legit Adyen webhooks are POST-only; GETs here are a probe
Related open alerts
If any open alerts match this path directly (e.g. a path_spike alert) OR any of the ASNs in the Top ASNs table (e.g. asn_spike on one of them), they appear here.
This answers "is a known bad ASN hitting a known priority path?" without the operator having to cross-check manually.
Pin, Ignore, Investigate IPs
Three actions on the page header:
- Pin — float this path to the top of the Paths list
- Ignore — hide from the list by default
- Investigate IPs — jumps to
/admin/ip-discoverywith this path pre-filled as the sampling filter. Use when the continuous Top IPs view shows—because volume is low but you suspect real abuse — the forensic sampler captures the full tuple list.
Low-volume path banner (added 2026-04-24)
If a path has request totals showing in the headline tiles but the Top ASNs / Cache / UA / HTTP methods panels below are all empty, you'll see an amber "Low-volume path — limited drill-down detail" banner above them.
Why: the headline counts and status-class timeseries come from the edge_path_totals AE dataset which has full path coverage. The dimensional panels query the main edge_requests rollup which keeps only the top-120 (path × ASN × country × UA × …) tuples per batch. A path with low per-tuple volume (say, 950 requests/hour spread across many ASNs) doesn't survive that cut, so no dimensional breakdown was captured.
What to do about it:
- Widen the range — 7d or 30d sums across many more batches and may surface dimensional data
- Use
/ips— the IP-level rollup uses a separate dataset and may capture activity for this path - Spin up an IP discovery window via the Investigate IPs button — forensic sampling captures the full tuple list at high resolution
The banner is silent on paths where dimensional data is healthy.