Proof of concept

koraiPulse — know where the slowness comes from, before your users do

When someone says “it’s slow”, koraiPulse helps you answer fast: the service, the server or the network. It checks your services from the outside, tracks the health of your servers and keeps the history in PostgreSQL.

Built on koraiCoreSelf-hostableYour data stays with you

What koraiPulse does today

Monitors of several kinds

Each check runs away from the monitored machine, from the locations you choose. A monitor is only declared down when most of its locations fail, after the number of retries you set.

  • HTTP: expected statuses, keyword present or absent, value at a JSON path
  • TCP, ping and DNS
  • Docker or Podman container started and healthy
  • Push: the monitored job calls its secret address, koraiPulse notices the silence
Monitors of every kind in one list, with availability over three periods.
A monitor’s response times, and each failure with the location that saw it.

Availability, maintenance and status pages

Availability reads over 24 hours, 30 days and one year. Maintenance windows are declared on screen. A public status page shows each monitor’s name and state and the published incidents, never its address or error messages.

A public status page: published incidents and each monitor’s history, with no address or error message.
A weekly maintenance window, limited to the monitors it concerns.

Server health, by the second and over time

The agent installed on each host reads its metrics every second and summarises them every minute: CPU, memory, disks and resource pressure. koraiPulse charts them, finds what runs on the machine and flags anomalies.

A week of disk activity on one server: the nightly batch shows.
What runs across the fleet, program by program.

Latency seen from the entrance of your sites

koraiPulse reads the korridor log and derives latency, throughput and errors minute by minute. Data stays for 45 days by the minute and 13 months by the hour.

A front-end over 24 hours: latency, traffic and the machine on one screen.

Your own indicators

An application can declare its own series and watches. koraiPulse stores and judges them without knowing the business, next to everything else.

Application watches, judged by koraiPulse without knowing their business.

Alerts that go through koraiCore

koraiPulse observes; koraiCore decides to alert and notifies, with its rules, channels and escalations. koraiPulse also inherits multi-tenancy and access rights from koraiCore.

On a phone: monitor state and the situation, judged by the koraiCore rules.

Frequently asked questions

Does koraiPulse need a time-series database?

No. koraiPulse relies on what the platform already produces, the korridor log and the koraiCore agent, and stores aggregated data in PostgreSQL.

Can I try koraiPulse?

Yes. An online demo presents a fictional fleet with its servers, sites and incidents. You can get the link from this page, and it holds no real data.

The screenshots on this page and the online demo run on fictional, deliberately small datasets: they illustrate only part of what the product does, and some features are limited or switched off in the demo environment.