From 987eae0408156e73540fef5d3b85f72309607281 Mon Sep 17 00:00:00 2001
From: Graham McIntire <graham@mcintire.me>
Date: Thu, 5 Feb 2026 09:25:29 -0600
Subject: [PATCH] fix: improve agent-based live polling with better logging and
 validation

- Added Logger.require for debug logging
- Refactored fetch_latest_agent_data to reduce nesting depth
- Added debug logging to trace agent assignment detection
- Check for nil/stale sensor readings and log appropriately
- Use actual reading timestamp instead of current time for data points

This will help diagnose why live polling shows 0 values and why
Phoenix SNMP disabled messages appear for agent-assigned devices.
---
 lib/towerops_web/live/graph_live/show.ex | 61 +++++++++++++++++-------
 1 file changed, 44 insertions(+), 17 deletions(-)

diff --git a/lib/towerops_web/live/graph_live/show.ex b/lib/towerops_web/live/graph_live/show.ex
index 8fe8d705..ef8c25a5 100644
--- a/lib/towerops_web/live/graph_live/show.ex
+++ b/lib/towerops_web/live/graph_live/show.ex
@@ -7,6 +7,8 @@ defmodule ToweropsWeb.GraphLive.Show do
   alias Towerops.Monitoring
   alias Towerops.Snmp
 
+  require Logger
+
   # Maximum live data points (5 minutes at 1 second intervals)
   @max_live_points 300
 
@@ -667,12 +669,18 @@ defmodule ToweropsWeb.GraphLive.Show do
     timestamp_ms = DateTime.to_unix(now, :millisecond)
 
     # Check if device has agent assignment
+    has_agent = socket.assigns[:has_agent_assignment] || false
+
+    Logger.debug("Live poll for device #{device.name}: has_agent=#{has_agent}")
+
     new_points =
-      if socket.assigns.has_agent_assignment do
+      if has_agent do
         # Use agent-collected data from database
+        Logger.debug("Using agent-collected data for live polling")
         fetch_latest_agent_data(socket.assigns, timestamp_ms)
       else
         # Do direct SNMP polling
+        Logger.debug("Using direct SNMP for live polling")
         client_opts = build_snmp_client_opts(device)
         poll_sensors_for_live_mode(socket.assigns, client_opts, timestamp_ms)
       end
@@ -696,28 +704,47 @@ defmodule ToweropsWeb.GraphLive.Show do
   end
 
   # Fetch latest sensor data collected by agent
-  defp fetch_latest_agent_data(assigns, timestamp_ms) do
+  defp fetch_latest_agent_data(assigns, _timestamp_ms) do
     sensors = get_sensors_for_live_mode(assigns)
 
     sensors
-    |> Enum.map(fn sensor ->
-      # Get the most recent sensor reading from the database
-      case Snmp.get_latest_sensor_reading(sensor.id) do
-        nil ->
-          nil
-
-        reading ->
-          %{
-            sensor_id: sensor.id,
-            label: sensor.sensor_descr,
-            value: Float.round(reading.value || 0.0, 1),
-            timestamp: timestamp_ms
-          }
-      end
-    end)
+    |> Enum.map(&fetch_sensor_reading/1)
     |> Enum.reject(&is_nil/1)
   end
 
+  defp fetch_sensor_reading(sensor) do
+    case Snmp.get_latest_sensor_reading(sensor.id) do
+      nil ->
+        Logger.debug("No reading found for sensor #{sensor.id} (#{sensor.sensor_descr})")
+        nil
+
+      %{value: nil} = reading ->
+        Logger.debug("Nil value for sensor #{sensor.id} (#{sensor.sensor_descr}) at #{reading.checked_at}")
+        nil
+
+      reading ->
+        validate_and_format_reading(sensor, reading)
+    end
+  end
+
+  defp validate_and_format_reading(sensor, reading) do
+    age_seconds = DateTime.diff(DateTime.utc_now(), reading.checked_at, :second)
+
+    if age_seconds > 300 do
+      Logger.debug("Stale reading for sensor #{sensor.id} (#{sensor.sensor_descr}), age: #{age_seconds}s")
+      nil
+    else
+      reading_timestamp_ms = DateTime.to_unix(reading.checked_at, :millisecond)
+
+      %{
+        sensor_id: sensor.id,
+        label: sensor.sensor_descr,
+        value: Float.round(reading.value, 1),
+        timestamp: reading_timestamp_ms
+      }
+    end
+  end
+
   # Build SNMP client options
   defp build_snmp_client_opts(device) do
     snmp_config = Devices.get_snmp_config(device)