Terms & Glossary

The vocabulary an agency meets across DeadWatch (monitoring) and Kestrel (report-selling) — one tight definition each, so a word on the board or in a report is never a mystery.

Terms are grouped by where you meet them. For the full picture of any capability, see the monitoring reference or the Kestrel diagnostics reference.

Monitoring & incidents

Uptime check — A scheduled request to a site that records response time, HTTP status, and a classified reason code, resolving to UP, DEGRADED, or DOWN.

Status (site) — A site's current state: UP (healthy), DEGRADED (responding but slow), DOWN (failing), UNKNOWN (not yet checked), or PAUSED (monitoring suspended).

Reason code — The machine-readable why behind a check result — e.g. curl_timeout, curl_ssl, http_503, success_slow — so a failure explains itself.

Incident — A confirmed problem grouped from start to resolution, moving through open → acknowledged → resolved. One incident represents one outage, however many failed checks it spans.

Severity — An incident's weight: major (from a DOWN condition) or minor (from a DEGRADED condition).

Confirmation re-check — The second (or third) failing check required before an incident opens, so a one-off blip doesn't page you but a real outage does. Server errors (HTTP 5xx) are trusted immediately.

Clean passes — The number of consecutive good checks required to resolve an incident (default 2).

Flapping — A site changing state too frequently (default: 3+ changes in 10 minutes). While flapping, new incidents are suppressed so one thrashing site produces one signal, not a storm.

Maintenance window — A scheduled period, one-off or recurring, during which checks still run but incidents are suppressed — so planned work doesn't page the on-call.

Short outage — A confirmed failure that resolved faster than the minimum-duration gate (default 45s). Counted, but not escalated as a full incident.

Health score — A single 0–100 number per site, weighted across pillars (availability, SSL, DNS, SEO, content, performance, links, VRT, security, freshness), that answers "how is this site overall" at a glance.

Pillar — One weighted component of the health score (e.g. Availability, worth 13; SSL, worth 12), each fed by its monitor.

Notification channel — Where an alert is delivered: Email, Slack, Discord, or a generic Webhook. Slack and Discord are auto-detected from their webhook URLs.

Notification routing — The rules that match an event type (e.g. uptime.down, escalation.start) to the channels that should receive it.

Visual, content & change monitors

Baseline — The approved reference for a page against which change is measured. In VRT, it's the accepted screenshot; a difference from it is what's reported.

VRT (Visual Regression Testing) — Screenshotting each page and comparing it to its baseline, per device breakpoint, to catch layout breaks a status check never would.

Diff % (diff_percent) — The share of a page that changed versus its VRT baseline. Lower is better; a spike after a deploy means something rendered differently than intended.

Breakpoint — A device viewport (mobile / tablet / desktop preset) that VRT and responsive checks run against, so a mobile-only break is still caught.

Drift — Change over time in a page's content or structure. Content drift hashes the HTML fingerprint (title, headings, meta, counts) and flags when the hash moves.

Link health — A score summarizing whether a page's links resolve — classified healthy (2xx), broken (4xx), or redirected/errored.

Asset change — A change in the JS/CSS/image assets a page loads, caught by fingerprinting the asset set into a snapshot hash — a deploy log and an early tamper signal in one.

Web Vitals & performance

Web Vitals — Google's user-experience metrics for a page's load and stability, measured in a real browser.

LCP (Largest Contentful Paint) — How long until the largest visible element renders; the headline "does it feel loaded" metric.

FCP (First Contentful Paint) — How long until the first content paints.

CLS (Cumulative Layout Shift) — How much the layout jumps around as it loads; lower is steadier.

TBT (Total Blocking Time) — How long the main thread was blocked, standing in for pre-interactivity jank.

TTFB (Time To First Byte) — How long until the server sends the first byte; a server-speed signal.

INP (Interaction to Next Paint) — How responsive the page is to user input.

CrUX (Chrome UX Report) — Real-user field data on Web Vitals, when Google has enough traffic for the site. Ranks above lab data for fidelity.

PSI (PageSpeed Insights) — Google's lab performance scoring (mobile and desktop), used when configured.

Lighthouse — The lab engine behind performance and accessibility scoring; a controlled, repeatable measurement.

Grade — A letter (A–F) mapping a 0–100 score to plain language — e.g. performance A ≥ 90 is "Fast," F is "Slow."

Kestrel & reports

Scan / probe — A grounded diagnostic run against a prospect's live site through the Scout engine; every finding traces to a real check, never a guess.

Finding — One observed issue, carrying a severity (critical / high / medium / low / info) and a phase.

Phase — A finding's priority band — urgent, hardening, or baseline — so the report reads as a prioritized worklist.

"Not checked" — A signal the scan could not verify (e.g. no CrUX field data). It is reported neutrally and never lowers the score — the fidelity principle that keeps a report honest.

Soft-404 — A page that returns HTTP 200 for URLs that don't exist. Kestrel detects catch-all hosts with control probes and suppresses false "exposed file" findings when it finds one.

Benchmark — A recognized best-practice target a report guides a site toward. Reports are framed as a guide to benchmarks, not a pass/fail verdict.

Disclaimer — The plain-language note on every report that automated diagnostics are a guide, not a guarantee, and that critical findings should be verified. The honesty that makes the report trustworthy.

Facet — A scored dimension of the public self-check (security, speed, seo, mobile, content, trust, tech), each blended into an overall bucket (good / warn / bad).

Tier (prospect) — An outreach-priority rank (Tier 1 / 2 / 3) assigned to a discovered business by weighted lead scoring, with a suggested price range and outreach method.

Coupon — A single-use, 30-day, 100%-off code (prefixed FLWP-) minted after a paid report, good for one free re-audit of the same URL — the hook that earns the re-audit.

Delivery — The automated pipeline that renders a paid report to PDF, emails it as an attachment, and issues a signed download link, then marks the purchase delivered.

Platform & access

Instance / tenant — Your agency's own deployment on its own subdomain, with data in its own database, never mixed with another agency's.

Site hierarchy — Parent → child grouping of sites, so a primary property and its related sites organize together.

Team — A group that owns sites and members; a viewer sees only the sites in their teams — the basis for scoped client access.

Role — A member's access level: super_admin, admin, member, viewer, or api_token, each granting a set of granular capabilities.

Capability — A single granular permission string (e.g. sites:create, incidents:manage, notifications:read) that roles are built from.

Permission set — A reusable template of capabilities (Designer, Developer, QA/Tester, DevOps Engineer, Project Manager, …) that can be imported as a starting point.

Site cap / overage — The plan's included site count (50 on the Agency plan). Additional sites bill per-site rather than breaking or locking the account.

License — The entitlement that governs a plan's included sites, features, and export permissions.