Zunnesco Monitor

Detect outages before customers report them, verify critical business journeys, alert the right person, and publish transparent service status from one platform.

This tree is Core 1.0.0. Author: zunnesco-dev. Plugin URI: https://github.com/teray234/zunnesco-monitor

The authoritative monitoring control plane is Laravel with geographically distributed probe workers. The WordPress client consumes signed APIs when a control plane URL is configured; otherwise it runs local scheduler / sandbox probes labeled as a single region — not a global SLA.

Core vs later editions

Next module in the original suite spec is Files, but Zunnesco Files already exists in this suite. Do not rebuild Files here.

Delayed on purpose: autonomous remediation, log ingestion, telephone on-call, private network probes, Playwright browser journeys. Synthetics are stub definitions only and never execute real purchases or production emails.

Local WordPress checks vs distributed Laravel probes

• WordPress only: WP-Cron runs HTTP and related checks as region sandbox. Useful for a single site. Results must not be advertised as multi-region uptime.
• Laravel control plane + probes/http: Probe workers poll signed assignments, validate SSRF, execute the check, sign the result, and POST to /api/v1/probes/results. A local worker can run as us-central.
• When control_plane_url is set in WP settings, the client consumes signed APIs instead of running the local scheduler.

Ownership

• Mail owns delivery; Monitor verifies and sends alert content through zunnesco_mail(), zunnesco_notifications(), or wp_mail.
• Domains owns registrar truth; Monitor watches expiration/DNS only if zunnesco_domains() exists.
• Credentials go through Vault when zunnesco_vault() exists; otherwise they are encrypted. Secrets are masked in logs.
• AI returns 501 unless zunnesco_ai() exists. AI must not claim a root cause without evidence and must not execute destructive remediation.

WordPress client

Activate Zunnesco Monitor. A homepage HTTP monitor is seeded from siteurl with a conservative 5-minute interval. Private IPs (loopback, metadata, RFC1918) are always blocked in Core.

Create an HTTP monitor

Admin → Monitor → Monitors, or:

$monitor = zunnesco_monitor()->monitors()->create([
    'type' => 'http',
    'name' => 'Commerce Checkout',
    'target' => 'https://example.com/checkout',
    'interval' => 60,
    'regions' => ['us-central'],
    'assertions' => [['type' => 'status', 'value' => 200]],
]);

Confirm an outage

A single failed check is recorded as unconfirmed (degraded). After N consecutive failures (default 3: initial → retry → sandbox second check) an incident opens and alerts fire. Acknowledge from Incidents (cap zunnesco_monitor_ack) or POST /wp-json/zunnesco-monitor/v1/incidents/{id}/acknowledge. Recovery requires N consecutive ups.

Publish a status page

Monitor → Status pages, map components to monitors, then embed [zunnesco_status_page]. Public pages are anonymous; private pages require login. Overall states: operational, degraded, partial, major, maintenance.

Send a heartbeat

zunnesco_monitor()->heartbeats()->complete($monitorId, ['duration_ms' => 1432]);

HTTP heartbeats require X-ZNMO-Signature and X-ZNMO-Timestamp (HMAC of timestamp.heartbeat:{token}).

REST: POST /wp-json/zunnesco-monitor/v1/heartbeats/{token}/complete

Private IPs

SSRF runs before every check. http://127.0.0.1/ and http://169.254.169.254/ are blocked. Core never allows private destinations (zunnesco_monitor/ssrf_allow_private is ignored in Core).

Laravel control plane

cd control-plane
cp .env.example .env
php artisan key:generate
touch database/database.sqlite
php artisan migrate --seed
php artisan serve
php artisan monitor:probe --region=us-central

Or run the CLI worker:

php probes/http/worker.php --url=https://example.com --region=us-central
php probes/http/worker.php --control-plane=http://127.0.0.1:8000 --probe-id=us-central-1 --secret=changeme

API (Bearer MONITOR_API_KEY except heartbeats and probe ingest):

POST /api/v1/monitors
GET  /api/v1/monitors/{id}
GET  /api/v1/monitors/{id}/checks
POST /api/v1/heartbeats/{token}/start|complete|fail
GET  /api/v1/incidents
POST /api/v1/incidents/{id}/acknowledge
POST /api/v1/maintenance-windows
GET  /api/v1/status-pages
POST /api/v1/probes/results   (signed probe)

Tests

php tests/test-schema.php

Caps, hooks

• Caps: zunnesco_monitor_manage, zunnesco_monitor_ack
• Filters: zunnesco_monitor/ssrf_allow_private, zunnesco_monitor/can_alert
• Actions: zunnesco_monitor/loaded, zunnesco_monitor/incident_opened, zunnesco_monitor/incident_resolved

SLA reports always expose interval, confirmation policy, maintenance treatment, and unknown-state handling. They never display 100% when unknown gaps exist.