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..fd0f0f6c428db4ef35c16a6c7e9f2286afd8b7a7
--- /dev/null
+++ b/infra/services/sysmon/android_device_metrics.py
@@ -0,0 +1,140 @@
+# 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 contextlib
+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

[ghost stip (do not use), 2016/05/16 07:39:16]

not used right now, but version bumps + cipd is a nightmare unless you have
rollforward/rollback capability

+ANDROID_DEVICE_FILE = os.path.join(os.path.expanduser('~'),

[ghost stip (do not use), 2016/05/16 07:39:16]

as of https://chromereviews.googleplex.com/426027013/ this file is dropped on
every android swarming bot

+                                   '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',

[ghost stip (do not use), 2016/05/16 07:39:16]

I tried to match these to
https://chromium.googlesource.com/chromium/tools/build/+/d648ab88761f198c77ac...

+                              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(
+    'dev/android_device_metric_read/status',
+    description='status of the last metric read')
+
+
+@contextlib.contextmanager
+def device_target():
+  """Switches metric transmission to the device target."""
+  old_target = interface.state.target
+
+  # From https://goo.gl/WuJDSu.

[M-A Ruel, 2016/05/16 23:01:13]

not a fan of using url shortener

[ghost stip (do not use), 2016/05/17 22:32:12]

Done.

+  fqdn = socket.getfqdn().lower()  # foo-[a|m]N.[chrome|golo].chromium.org
+  host = fqdn.split('.')[0]  # foo-[a|m]N
+  region = ts_mon.config._default_region(fqdn)
+  network = ts_mon.config._default_network(host)
+
+  interface.state.target = targets.DeviceTarget(

[ghost stip (do not use), 2016/05/16 07:39:16]

I don't know if you can just *do* this, leaving it up to dsansome@ and
sergeyberezin@

[Sergey Berezin, 2016/05/16 19:01:43]

Nope, this is effectively a dangerous no-op. The default target is used at the
time of metric flush, which runs in an async thread, and what you're doing is
not thread-safe.

Instead, if you really want to change the target fields, use target_fields={...}
in the metric's .set() method.

But you probably don't want to modify target fields anyway, because why would
you? You already have your own metric names. I don't think we rely on this value
in the old metrics. The only reason it was there was to avoid clashing with
sysmon on presence/up metric. But now that we send it from sysmon itself, there
is no clash. So, I'd say don't modify the target fields at all.

[ghost stip (do not use), 2016/05/17 22:32:12]

Done.

+      region,
+      'temperature_monitor',
+      network,
+      host)
+
+  yield
+
+  interface.state.target = old_target
+
+
+def get_device_statuses():
+  success, devices = _load_android_device_file()
+  if not success:
+    return
+
+  with device_target():
+    for device_name, device in devices.iteritems():
+      fields = {'device_id': device_name}
+
+      # Fields with special handling.
+      cpu_temps = device.get('temp', [])
+
+      battery_temp = device.get('battery', {}).get('temperature')
+      battery_temp = battery_temp / 10.0 if battery_temp else None
+
+      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),

[ghost stip (do not use), 2016/05/16 07:42:30]

maruel: I'm hoping that swarming is smart enough in
https://goto.google.com/cyjvj to report the 'useful' temperature sensor first.
if that isn't the case, I'd like to export the blacklist out from swarming
somehow

[M-A Ruel, 2016/05/16 23:01:13]

Update the bot_config as you need if you find anything wrong.

[ghost stip (do not use), 2016/05/17 22:32:12]

Acknowledged.

+          (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:

[ghost stip (do not use), 2016/05/16 07:39:16]

I feel like I'm missing another Error here... maybe for windows?

[M-A Ruel, 2016/05/16 23:01:13]

(IOError, OSError)

[ghost stip (do not use), 2016/05/17 22:32:13]

Done.

+    # 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', [])