tj@siedsma:~/apps/iceberg$ DSM, CDT
$ cd ~/apps

IceBerg//Monitoring

Monitoring for everything I run.

IceBerg is the monitoring app I built to know about issues before they become problems. Like an iceberg, the small part you can see warns you long before you hit what's under the water, and being in the know is what matters.

It watches the websites, servers and scheduled jobs behind my products and client sites: HTTP uptime, TLS certificates, DNS, domain registrations, background jobs and server health.

Monitoring, alerting and status reporting all run on infrastructure I control.

tj@siedsma:~$ iceberg status

Already on it.

IceBerg status page showing all systems operational with 90 day uptime bars for five sites
Simulated data, real domain names.

Redundancy

Two nodes on separate providers.

The primary runs every check, sends the alerts and serves the public status pages. A standby on a second network watches the primary and pages me if it ever goes quiet, so a monitoring outage never goes unnoticed.

Diagram of the IceBerg primary node and its standby on a separate provider

Coverage

Jobs and servers, not just sites.

Floe, a lightweight signed agent, reports load, memory, disk and database health from each server without credentials ever leaving the machine. My Laravel applications report their scheduled jobs through a heartbeat package I publish. Certificate and domain expiry are tracked for every site I run.

IceBerg page for a server running Floe, with its monitors and latest load, memory, disk, Redis and MySQL figures

Alerting

Pages until someone owns it.

Critical alerts keep paging until someone acknowledges the incident in IceBerg, not just on their phone. Certificates and domains send reminders well before they expire instead of raising false outages. Public status pages give clients an honest view of service health, with maintenance labeled separately from outages.

IceBerg incident for a booking checkout outage, acknowledged after six minutes, with the failed checks listed

AI assistants

Ask it in plain language.

Through its MCP server and a scoped API, I can ask what is down, acknowledge an incident or start monitoring a new site, and every change is logged with where it came from.

Claude answering which domains expire in the next 45 days by checking IceBerg: one, with 38 days left