Arbitrary Systems Ops
A private operator surface for monitoring, troubleshooting, and safely repairing The Registry, PranaLogic, I'm open 2, and Group Pours as they move out of beta.
Live monitoring is intentionally not implied yet. The next milestone is adding signed health adapters to each product.
Product Family
These cards are the first shared registry for what operators need to watch. Live checks, deploy metadata, and repair actions will attach here as app adapters come online.
Registry
Collector-first inventory and agent-ready workflows for wine, spirits, beer, cigars, and related collections.
Watch imports, catalog cleanup, billing/webhooks, Group Pours sync, and agent workflow contracts.
1 initial runbook queued
Prana
Studio operating system for scheduling, customers, waivers, memberships, staff, payments, and owner repair tools.
Watch booking flow, payments, reminders, waivers, customer records, Twilio, Resend, and platform-admin repairs.
1 initial runbook queued
IOT
Private, circle-based social planning around Pulses, Echos, Syncs, invites, and agent access.
Watch production smoke checks, invites, circles, Pulse flow, billing, email delivery, and agent/MCP surfaces.
1 initial runbook queued
GP
Collaborative tasting and group note-taking product connected to Registry-style collection workflows.
Watch tasting sessions, share codes, GPAT notes, feedback, user tools, and Registry push failures.
1 initial runbook queued
Build Path
The safest first move is central visibility and written runbooks. Mutation comes later, behind signed app-owned endpoints and audit trails.
Private shell, app registry, scope docs, roles, incidents, runbooks, and audit model.
Scheduled checks, normalized smoke output, uptime, latency, and app-owned health adapters.
Searchable diagnostics, failure queues, external service status, and deploy timelines.
Guarded, audited, app-owned repair actions with confirmations and escalation guidance.
Initial Runbooks
Trigger: A collector import, catalog cleanup, or Group Pours push reports a sustained failure.
First action: Check the affected object, confirm whether the failure is isolated or systemic, and avoid rerunning backfills until the duplicate-risk is understood.
Escalate: Escalate to engineering if the same catalog item, webhook event, or sync job fails twice after a controlled retry.
Trigger: A studio cannot book, cancel, charge, or reconcile a customer-facing transaction.
First action: Check API health, Stripe webhook status, and platform-admin customer timeline before applying any repair action.
Escalate: Escalate immediately if multiple studios are affected or if money moved without a matching booking/payment record.
Trigger: Production smoke check, invite flow, or agent endpoint starts failing.
First action: Run the production smoke check, inspect the failing route category, and verify whether the issue affects public pages, app pages, or agent-only surfaces.
Escalate: Escalate if guarded endpoints expose unexpected data, if invites fail broadly, or if smoke failures persist after one redeploy.
Trigger: A live tasting cannot be joined, notes cannot be recorded, or Registry push fails.
First action: Look up the share code or response ID, inspect Registry push status and last error, then determine whether the tasting can continue without sync.
Escalate: Escalate if active participants are blocked or if completed tasting notes are at risk of being lost.