Optional mixin for a Monitor that reacts to external check-in pings — Raven's /checkin/:checkin_id HTTP endpoint — in addition to (not instead of) its own periodic collect/2.

This is a narrower, different shape than CodeNameRaven.Probe: Probe fully replaces collect/2 with a reactive stub, for a monitor whose only purpose is watching another monitor's events. A check-in monitor still runs a real, meaningful collect/2 on every scheduled interval (typically a staleness check against its own last-check-in timestamp) — it just also needs an event-driven hook to update that timestamp when a ping arrives. Layer this alongside a normal use CodeNameRaven.Monitor, not in place of it.

Like CodeNameRaven.Probe, this module is implementation-free: it defines only the callback contract. Raven itself subscribes a check-in-shaped monitor to check-in events on its behalf and calls on_checkin/3 directly — the module never subscribes to anything itself and has no dependency on any Raven-internal module.

Using this behaviour

defmodule Integrations.MyDeadMansSwitch do
  use CodeNameRaven.Monitor
  use CodeNameRaven.Monitor.Checkin

  @impl true
  def on_checkin(%{checkin_id: incoming_id}, params, state) do
    if incoming_id == params[:checkin_id] do
      {:ok, %{state | last_checkin_at: DateTime.utc_now()}, :check_now}
    else
      {:ok, state}
    end
  end

  @impl true
  def collect(params, state) do
    # staleness check against state.last_checkin_at
  end

  @impl true
  def healthy?(result), do: # :up / :degraded / :down based on elapsed time
end

Event shape

on_checkin/3's first argument is loosely typed (term()) rather than a concrete struct, matching CodeNameRaven.Probe's on_metric/3/ on_status/3 precedent. It has (at least) a checkin_id field (string — the id from the check-in URL) and a checkin_at field (DateTime.t()).