pulsefy.tools

Generic event monitoring & alerting SaaS. Wire any app event to Telegram / Slack / webhook via rule-based thresholds.

Deployment

Runs on DigitalOcean droplet as systemd service pulsefy on 127.0.0.1:8093, fronted by nginx on pulsefy.tools.

Reads events read-only from searchads.tools events database (/root/.friday-asa/asa.db) — production shares that ingestion pipeline for now. Own SQLite for users / sessions / apps: /opt/pulsefy/state/pulsefy.db.

Structure

• server.js — HTTP + rule engine + SSE (dependency-free, node:http)
• public/ — landing.html, auth.html, onboarding.html, app.html
• config/ — rules.json, integrations.json (NOT tracked)
• state/ — SQLite users/sessions/apps + alarms.jsonl (NOT tracked)

Multi-tenancy

Each user claims apps via application_token (the same UUID indie-apps send to the events ingest). Cross-checked against asa_accounts.app_token in the events DB. Events queries filter by user's claimed asa_account_ids.

Routes

• / — landing (public) or 302 → /app (auth)
• /login /register — auth form
• /onboarding — first-app claim (auth, no apps)
• /app — dashboard (auth, ≥1 app)
• /api/auth/* — register / login / logout / me
• /api/apps — CRUD
• /api/events — filtered feed
• /api/events/live — SSE (filtered)
• /api/events/summary — event count breakdown
• /api/rules — CRUD
• /api/integrations — telegram / slack / webhook config
• /api/alarms — fired alarm history

Rule engine

Runs every 60s. Per rule:

• events + window + threshold + severity
• optional filter_props.error_contains
• cooldown per rule-signature (deduped by first ~60 chars of error message)
• fans out to configured channels