Overall architecture

Knock daemon is build around three components :

Diagram:

Knock Daemon handles probes locally on your server(s) and pushes results asynchronously to your managed namespace.

Transport uses json/gzip over https.

Our daemon embeds a localhost UDP listener (port 63184) which handles application counters coming from SDKs.

We send mainly server hostname and probes datas toward our infrastructure (integers/floats, some string buffers - for inventory for instance).

The managed supervision namespace provides a secured web dashboard and fires notifications.

Knock Daemon is backed by an upgrade cron which keep it up-to-date:

  • Cron logs are located inside "/var/log/knock-autoupdate.log".
  • Cron is enable Monday to Tuesday (included), for period 08h-18h (local server), once per hour.

Account, namespace and users

The account you create when you sign-up has:

  • An email address, which must be validated.
  • A password, which is used to login to web site.
  • A company name, or namespace. It identifies your organisation and is required for api calls.
  • A namespace api key, private to you, and generated during sign-up.
  • Two emails addresses for notifications (refer to : How notifications behave).

You can create additionnal users. These users will only have access to supervision web frontend.

Servers handling

By design, servers are identified by their hostname.

Having several servers with same hostname is not supported (it will work, but probe results will be inconsistent).

Renaming a server is supported, new datas will just target he new host name.

In order to notify your supervision namespace that a server is now shutdown / in maintenance mode, a JSON Api end-point will be available soon.

Active server

We bill on a server basis, with per hour pricing.

An active server - identified by its hostname - is a server for which we received at least one probe within one hour period.

Knock managed namespace

We manage your supervision namespace.

This includes :

  • Existing probes, graphs, alerts and notifications upgrades.
  • New probes, graphs, alerts and notifications deployment.