# Equinix — IBX data center pages + global press-release feed
id: equinix
name: Equinix
domain: www.equinix.com
kind: operator
mode: hybrid
priority: 1
license: "Publicly accessible operator pages; factual data only; © Equinix, Inc."
attribution: "Equinix"
homepage: https://www.equinix.com/data-centers
notes: >
  www.equinix.com is behind Akamai: the bot user-agent gets HTTP 403, a browser identity (level 2, still direct HTTP)
  is served normally. Level 1 is kept as the start so the runtime escalates only when blocked; Scrapfly (L4) is the
  last resort and is not reached for IBX pages (0 credits in dry runs). Per-IBX pages publish address + coordinates +
  colocation space + certifications, but no MW figure — none is invented.
  NEWSROOM DISABLED (2026-09-11): press releases live on newsroom.equinix.com (RSS
  https://newsroom.equinix.com/press-releases-global?pagetemplate=rss, URLs /YYYY-MM-DD-<slug>). Those pages are
  served normally (HTTP 200, ~20 KB text) but embed Cloudflare's "/cdn-cgi/challenge-platform/" beacon, which the
  shared looksBlockedOrEmpty() heuristic treats as a bot wall → every article escalated to Scrapfly (30 credits each,
  120 credits in the first dry run). Until the heuristic is tightened (only treat the marker as blocked on short
  pages), the feed is left out rather than burning premium credits on open pages.
coverage: { global: true, operators: ["Equinix"] }
fetch:
  level: 1
  maxLevel: 4
  rpm: 20
  concurrency: 2
  respectRobots: true
  maxCreditsPerRun: 60
  maxCreditsPerDay: 120   # premium credits per UTC day for this connector (all runs, all workers)
discovery:
  sitemap: ["/sitemap-core.xml"]
  rss: ["https://newsroom.equinix.com/press-releases-global?pagetemplate=rss"]   # fetcher heuristic fixed 2026-09-11: challenge-platform beacon on content-rich pages no longer escalates
  include:
    - "^https://www\\.equinix\\.com/data-centers/"
    - "^https://newsroom\\.equinix\\.com/"
  exclude:
    - "\\?"
    - "/data-centers/(colocation|design|data-center-excellence|services|solutions)"
  classify:
    - { pattern: "/data-centers/[a-z-]+-colocation/[a-z-]+-colocation/[a-z0-9-]+-data-centers/[a-z]{2,3}\\d{1,2}[a-z]?$", group: facility_pages, pageType: facility_page, priority: 60 }
    - { pattern: "/data-centers/[a-z-]+-colocation/[a-z-]+-colocation/[a-z0-9-]+-data-centers$", group: index, pageType: facility_index, priority: 20 }
    - { pattern: "/data-centers/[a-z-]+-colocation(/[a-z-]+-colocation)?$", group: index, pageType: facility_index, priority: 10 }
    - { pattern: "/data-centers/", group: index, pageType: facility_index, priority: 5 }
  maxUrlsPerRun: 2000
schedule:
  facility_pages: weekly
  newsroom: daily
  index: monthly
  sitemap: weekly
  rss: daily
extractors:
  facility:
    # selected by URL only: classifyPage() may mislabel deep facility URLs (e.g. "closure"), the URL is authoritative
    match: ["/data-centers/[a-z-]+-colocation/[a-z-]+-colocation/[a-z0-9-]+-data-centers/[a-z]{2,3}\\d{1,2}[a-z]?$"]
    kind: facility
    parser: equinix_facility_v1
defaults:
  operatorName: Equinix
  operatorKind: colocation
  facilityType: colocation
parserVersion: v1