MonSPHERE

Monitoring

Port Discovery

Full-range TCP port scan of a host you own to flag any port that's open now but wasn't before — an early warning for an accidentally-exposed service.

A Port Discovery monitor scans a host's entire TCP port range (0-65535) on every check and reports any port that's open now but wasn't on the previous check — catching an accidentally-exposed database, admin panel, or debug port before it becomes a real incident. Only scan infrastructure you actually own or have explicit authorization to test — scanning a target you don't control can be a legal or contractual problem regardless of intent.

Availability

Requires a plan with Port Discovery (Professional and above) — a full 65535-port scan is the single heaviest check this platform runs, with its own dedicated scanning infrastructure per check.

How it works

curl -X POST https://api.monsphere.com/api/v1/organizations/{orgId}/monitors \
  -H "X-Api-Key: msph_live_..." -H "Content-Type: application/json" \
  -d '{
    "name": "Production Host Port Scan",
    "type": "port-discovery",
    "target": "203.0.113.10",
    "intervalSeconds": 21600,
    "timeoutSeconds": 180,
    "config": { "portDiscoveryKnownPorts": [22, 443] }
  }'
  • portDiscoveryKnownPorts — optional. If you list the ports you already expect to be open, the very first scan can already alert on anything unexpected. Leave it out and the first scan establishes the baseline automatically instead, the same "first check never alerts" behavior every other change-detection monitor type uses — with the tradeoff that a port already (wrongly) open at monitor-creation time would be silently absorbed into that baseline with no chance to catch it.
  • A full scan takes noticeably longer than any other check type (well over a minute against a typical host) — the default interval is 6 hours and the default timeout 3 minutes; both are adjustable, but a full port scan isn't something worth running every 60 seconds.

How it runs

A scan this size doesn't fit the same fast in-process check every other monitor type uses — it runs on its own dedicated queue and worker process, isolated from every other monitor's checks so a multi-minute scan never delays anything else. There's no region to pick, same as WHOIS and Certificate Transparency.