Send ts_mon metrics from android device data.

infra/services/sysmon/android_device_metrics.py

Index: infra/services/sysmon/android_device_metrics.py
diff --git a/infra/services/sysmon/android_device_metrics.py b/infra/services/sysmon/android_device_metrics.py
new file mode 100644
index 0000000000000000000000000000000000000000..8883820936e003997835cd72c2d78589349ead5a
--- /dev/null
+++ b/infra/services/sysmon/android_device_metrics.py
@@ -0,0 +1,116 @@
+# Copyright (c) 2016 The Chromium Authors. All rights reserved.
+# Use of this source code is governed by a BSD-style license that can be
+# found in the LICENSE file.
+
+import json
+import logging
+import os
+import socket
+import time
+
+
+from infra_libs import ts_mon
+from infra_libs.ts_mon.common import interface
+from infra_libs.ts_mon.common import targets
+
+
+ANDROID_DEVICE_FILE_VERSION = 1
+
+ANDROID_PREVIOUS_DEVICE_FILE_VERSION = ANDROID_DEVICE_FILE_VERSION - 1
+ANDROID_DEVICE_FILE = os.path.join(os.path.expanduser('~'),
+                                   'android_device_status.json')
+
+# Don't read a file older than this many seconds.
+ANDROID_DEVICE_FILE_STALENESS_S = 120
+
+
+cpu_temp = ts_mon.FloatMetric('dev/cpu/temperature',
+                              description='device CPU temperature in deg C')
+batt_temp = ts_mon.FloatMetric('dev/battery/temperature',
+                               description='battery temperature in deg C')
+batt_charge = ts_mon.FloatMetric('dev/battery/charge',
+                                 description='percentage charge of battery')
+dev_status = ts_mon.StringMetric('dev/status',
+                                 description='operational state of device')
+dev_os = ts_mon.StringMetric('dev/os',
+                             description='operating system of the device')
+dev_uptime = ts_mon.FloatMetric('dev/uptime',
+                                description='device uptime in seconds')
+
+metric_read_status = ts_mon.StringMetric(

[bpastene, 2016/05/19 18:36:44]

What are your plans for this metric?

[ghost stip (do not use), 2016/05/20 21:12:50]

potential debugging, also a quick way to see if a bot is swarming or not

+    'dev/android_device_metric_read/status',
+    description='status of the last metric read')
+
+
+def get_device_statuses():
+  success, devices = _load_android_device_file()
+  if not success:

[bpastene, 2016/05/19 18:36:44]

Nit: Remove success from the return and replace this line with
'if not devices:' ? IDK; how do I python?

[ghost stip (do not use), 2016/05/20 21:12:50]

sure, why not

+    return
+
+  for device_name, device in devices.iteritems():
+    fields = {'device_id': device_name}
+
+    # Fields with special handling.
+    cpu_temps = device.get('temp', [])

[bpastene, 2016/05/19 18:36:44]

I don't think this is a list as reported by swarming. Looks like a dict to me. A
sample test file should pick this up.

[ghost stip (do not use), 2016/05/20 21:12:50]

you're right

+
+    battery_temp = device.get('battery', {}).get('temperature')
+    battery_temp = battery_temp / 10.0 if battery_temp else None

[bpastene, 2016/05/19 18:36:44]

I've been burned repeatedly with some devices reporting millicelsius, some
reporting decicelsius, and some reporting celsius. Hopefully bot_config
normalizes all this.

[ghost stip (do not use), 2016/05/20 21:12:50]

yikes

+
+    status = device.get('state')
+    status = 'good' if status == 'available' else status
+
+    for metric, value in (
+        (cpu_temp, cpu_temps[0] if cpu_temps else None),
+        (batt_temp, battery_temp),
+        (batt_charge, device.get('battery', {}).get('level')),
+        (dev_status, status),
+        (dev_os, device.get('build', {}).get('build.id')),
+        (dev_uptime, device.get('uptime', None))):
+      if value is not None:
+        metric.set(value, fields=fields)
+
+
+def _load_android_device_file():
+  """Load the android device file and check for errors or staleness."""
+  try:
+    with open(ANDROID_DEVICE_FILE) as f:
+      file_data = f.read()
+  except (IOError, OSError):
+    # File isn't there, not an Android bot.
+    metric_read_status.set('not_found')
+    logging.debug('Android device file %s not found', ANDROID_DEVICE_FILE)
+    return False, None
+
+  try:
+    json_data = json.loads(file_data)
+  except ValueError as e:
+    metric_read_status.set('invalid_json')
+    logging.error('Android device file %s invalid json: %s',
+                  ANDROID_DEVICE_FILE, e)
+    return False, None
+
+  if not isinstance(json_data, dict):
+    metric_read_status.set('invalid_json')
+    logging.error('Android device file %s is not a dict', ANDROID_DEVICE_FILE)
+    return False, None
+
+  if json_data.get('version') not in (
+      ANDROID_DEVICE_FILE_VERSION,
+      ANDROID_PREVIOUS_DEVICE_FILE_VERSION):
+    metric_read_status.set('invalid_version')
+    logging.error('Android device file %s is version %s, not %s',
+                  ANDROID_DEVICE_FILE, json_data.get('version'),
+                  ANDROID_DEVICE_FILE_VERSION)
+    return False, None
+
+  now = time.time()
+  timestamp = json_data.get('timestamp', 0)
+  if now >= timestamp + ANDROID_DEVICE_FILE_STALENESS_S:
+    metric_read_status.set('stale_file')
+    logging.error('Android device file %s is %ss stale (max %ss)',
+                  ANDROID_DEVICE_FILE, now - timestamp,
+                  ANDROID_DEVICE_FILE_STALENESS_S)
+    return False, None
+
+  metric_read_status.set('good')
+  return True, json_data.get('devices', [])