Graham McIntire
291ecd1054
Implements Icinga2-style generic check system for comprehensive monitoring. Phoenix Backend: - Generic checks table with JSONB config for all check types - TimescaleDB check_results hypertable for time-series data - Check executors (HTTP, TCP, DNS) - CheckWorker with self-scheduling and state transitions - Extended alerts system for check-based alerting Check Types: HTTP/HTTPS, TCP, DNS Architecture: Icinga2 Checkable pattern, soft/hard states, flapping detection Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
56 lines
1.6 KiB
Elixir
56 lines
1.6 KiB
Elixir
defmodule Towerops.Monitoring.Executor do
|
|
@moduledoc """
|
|
Dispatcher for executing checks based on check_type.
|
|
|
|
Routes to type-specific executors (HTTP, TCP, DNS, etc.).
|
|
"""
|
|
|
|
alias Towerops.Monitoring.Check
|
|
alias Towerops.Monitoring.Executors.DnsExecutor
|
|
alias Towerops.Monitoring.Executors.HttpExecutor
|
|
alias Towerops.Monitoring.Executors.TcpExecutor
|
|
|
|
@doc """
|
|
Executes a check and returns the result.
|
|
|
|
Returns:
|
|
- {:ok, response_time_ms, output} on success
|
|
- {:error, reason} on failure
|
|
"""
|
|
def execute(%Check{check_type: "http", config: config, timeout_ms: timeout}) do
|
|
HttpExecutor.execute(config, timeout)
|
|
end
|
|
|
|
def execute(%Check{check_type: "tcp", config: config, timeout_ms: timeout}) do
|
|
TcpExecutor.execute(config, timeout)
|
|
end
|
|
|
|
def execute(%Check{check_type: "dns", config: config, timeout_ms: timeout}) do
|
|
DnsExecutor.execute(config, timeout)
|
|
end
|
|
|
|
def execute(%Check{check_type: "ping"}) do
|
|
# Ping is handled by DeviceMonitorWorker, not here
|
|
{:error, "Ping checks are handled separately"}
|
|
end
|
|
|
|
def execute(%Check{check_type: "passive"}) do
|
|
# Passive checks don't execute - they receive results
|
|
{:error, "Passive checks do not execute actively"}
|
|
end
|
|
|
|
def execute(%Check{check_type: type}) do
|
|
{:error, "Unknown check type: #{type}"}
|
|
end
|
|
|
|
@doc """
|
|
Converts executor result to check status code.
|
|
|
|
Returns:
|
|
- 0 = OK
|
|
- 2 = CRITICAL (for errors)
|
|
- 3 = UNKNOWN (for unexpected errors)
|
|
"""
|
|
def result_to_status({:ok, _response_time, _output}), do: 0
|
|
def result_to_status({:error, _reason}), do: 2
|
|
end
|