Allow botlist API call to respond to quarantined: and is_dead:

appengine/swarming/server/bot_management.py

 # Copyright 2014 The LUCI Authors. All rights reserved.
 # Use of this source code is governed under the Apache License, Version 2.0
 # that can be found in the LICENSE file.
 
 """Swarming bot management, e.g. list of known bots and their state.
 
           +---------+
           |BotRoot  |
           |id=bot_id|
           +---------+
@@ skipping 10 matching lines @@
   happening for the bot.
 - BotInfo is a 'dump-only' entity used for UI, it permits quickly show the
   state of every bots in an single query. It is basically a cache of the last
   BotEvent and additionally updated on poll. It doesn't need to be updated in a
   transaction.
 - BotSettings contains bot-specific settings. It must be updated in a
   transaction and contains admin-provided settings, contrary to the other
   entities which are generated from data provided by the bot itself.
 """
 
+import datetime
 import hashlib
 
 from google.appengine.ext import ndb
 
+import swarming_rpcs
 from components import datastore_utils
 from components import utils
 from server import config
 from server import task_pack
 
 

[M-A Ruel, 2016/08/09 20:23:13]

remove one line

[kjlubick, 2016/08/09 20:44:44]

Done.

+
 # Margin of randomization of BOT_REBOOT_PERIOD_SECS. Per-bot period will be in
 # range [period * (1 - margin), period * (1 + margin)).
 BOT_REBOOT_PERIOD_RANDOMIZATION_MARGIN = 0.2
 
 
 ### Models.
 
 # There is one BotRoot entity per bot id. Multiple bots could run on a single
 # host, for example with multiple phones connected to a host. In this case, the
 # id is specific to each device acting as a bot.
@@ skipping 209 matching lines @@
   if order:
     query = query.order(BotEvent.key)
   return query
 
 
 def get_settings_key(bot_id):
   """Returns the BotSettings ndb.Key for a known bot."""
   return ndb.Key(BotSettings, 'settings', parent=get_root_key(bot_id))
 
 
+def filter_dimensions(q, dimensions):
+  """Filters a ndb.Query for BotInfo based on dimensions in the request."""
+  for d in dimensions:
+    parts = d.split(':', 1)
+    if len(parts) != 2 or any(i.strip() != i or not i for i in parts):
+      raise ValueError('Invalid dimensions')
+    q = q.filter(BotInfo.dimensions_flat == d)
+  return q
+
+
+def filter_availability(q, quarantined, is_dead, now):
+  """Filters a ndb.Query for BotInfo based on quarantined/is_dead."""
+  val = swarming_rpcs.to_bool(quarantined)

[M-A Ruel, 2016/08/09 20:23:13]

actually that's a layering violation :/  Here only python types should be used.

[kjlubick, 2016/08/09 20:44:44]

Done.

+  if val is not None:
+    q = q.filter(BotInfo.quarantined == val)
+
+  dt = datetime.timedelta(seconds=config.settings().bot_death_timeout_secs)
+  timeout = now - dt
+  val = swarming_rpcs.to_bool(is_dead)
+  if val:
+    q = q.filter(BotInfo.last_seen_ts < timeout)
+  elif val is not None:
+    q = q.filter(BotInfo.last_seen_ts > timeout)
+  return q
+
+
 def bot_event(
     event_type, bot_id, external_ip, authenticated_as, dimensions, state,
     version, quarantined, task_id, task_name, **kwargs):
   """Records when a bot has queried for work.
 
   Arguments:
   - event: event type.
   - bot_id: bot id.
   - external_ip: IP address as seen by the HTTP handler.
   - authenticated_as: bot identity as seen by the HTTP handler.
@@ skipping 88 matching lines @@
   Returns:
     Tuple (True to restart, text message explaining the reason).
   """
   # Periodically reboot bots to workaround OS level leaks (especially on Win).
   running_time = state.get('running_time', 0)
   assert isinstance(running_time, (int, float))
   period = get_bot_reboot_period(bot_id, state)
   if period and running_time > period:
     return True, 'Periodic reboot: running longer than %ds' % period
   return False, ''