PulseDeck is a single Python file that speaks Monit's own protocol, drops straight onto an existing M/Monit database, and gives you back a dashboard you can actually read at 3am — plus the SNMP, WinRM, and Prometheus reach that M/Monit never had.
No account. No phone-home telemetry. No Docker required — one install script, one Python file.
Built to replace, not orbit around, the tools you already run:
No roadmap slides. This is the current feature set, version 1.19.0.
Linear regression on disk usage forecasts when a volume will cross your threshold, not just that it already has. Link-state transition counting catches a flapping interface that never trips a normal threshold check. Ping-jitter variance flags an unstable connection even when the average latency looks perfectly fine.
Poll switches, Windows hosts, and any Prometheus exporter alongside your Monit agents — one dashboard, one event log, one alert pipeline.
Run PulseDeck at each site. Point one at another as a "satellite" and its hosts show up on a shared dashboard, auto-grouped by zone — without merging databases.
Custom {{host}} / {{message}} templates per channel, with automatic digest batching when a real outage tries to send you forty individual emails at once.
Full-width charts, threshold bands drawn right on the graph, drag-to-zoom, a latency heatmap, and event markers lined up against the exact metric spike that caused them.
test_pulsedeck.py runs 32 real checks against a live instance after every install or update — and cleans up every bit of test data it creates.
Filter the whole dashboard down to one zone, one rack, one team's hosts, with a link — not a spreadsheet you maintain by hand on the side.
English, German, Bulgarian, French, Dutch, Spanish — every string, every tab, every error message, not just the landing page.
An explicit opt-in allowlist, not "everything's public by default." Pick which hosts show up, give them a customer-facing name, and get a 90-day uptime history bar per component. No login required — nothing not explicitly added is ever reachable.
Roll up services from different hosts into one thing your team actually cares about — "is Checkout working" instead of five separate technical statuses. Worst-of-members: any broken piece makes the whole thing read as down.
Declare that your app depends on the database it talks to, and stop getting paged for the symptom instead of the cause. Now with a visual diagram of the whole chain, color-coded by live status — not just a list of IDs.
Disk forecasts and jitter detection catch specific, known conditions. This catches the ones nobody thought to write a rule for — a statistical baseline per metric, flagging anything that drifts too many standard deviations from its own normal.
Per-host uptime % over 7/30/90 days, worst hosts sorted first, one click to a PDF you can actually hand to someone asking for proof.
Drag panels into the order that matches how you triage, hide the ones you never look at. Saved per person, not forced on the whole team.
Not trying to replace Grafana or Zabbix — trying to replace the M/Monit box you resent maintaining.
| PulseDeck | M/Monit | Zabbix | Grafana | |
|---|---|---|---|---|
| Speaks the Monit agent protocol directly | ✓ | ✓ | – | – |
| Drops onto an existing M/Monit MySQL schema | ✓ | ✓ (it is the schema) | – | – |
| SNMP / WinRM / Prometheus polling built in | ✓ | – | ✓ | via data source |
| Predictive/trend-based alerting (disk forecast, flapping, jitter, statistical anomalies) | ✓ | – | partial | – |
| Public, no-login status page for external stakeholders | ✓ | – | – | partial — shares a raw dashboard, not a stakeholder-facing status page |
| Single-file deployment, no external services required | ✓ | – | – | – |
| Deep, general-purpose dashboarding across arbitrary data sources | not the goal | – | – | ✓ |
| License cost | Free | Paid, per-host | Free (self-hosted) | Free tier + paid |
Feature comparisons based on each product's publicly documented capabilities at time of writing. If something above looks wrong, we'd genuinely like to know.
No microservices to lose track of, no message queue to run out of disk, no separate scheduler process to forget you have. `git diff` on one file tells you the whole story of a change.
Your Monit agents don't need to know anything changed. Point them at PulseDeck's collector endpoint instead, and every check, event, and history row lands in a schema that isn't proprietary to us.
An offline install bundle ships Python and every dependency as wheels, for the boxes that will never see the public internet — which, if you're reading this, is probably most of them.
MIT-licensed. Runs anywhere Python 3.10+ and MySQL/MariaDB run. No telemetry, no phone-home, no account required to use every feature on this page.
.envThe dashboard itself will never be paywalled. What's coming later is optional — managed hosting, multi-tenant billing, and support contracts for teams who'd rather not run the box themselves.
Every feature on this page, forever.
We run it, patch it, and back it up. You just look at the dashboard.