Report processes that do not belong to the agent's cgroup #1908
Conversation
narrieta
requested review from
kevinclark19a,
larohra,
pgombar and
ZhidongPeng
as code owners
| def get_monitor_handler(): | ||
| return MonitorHandler() | ||
|
|
||
|
|
||
| class PollResourceUsageOperation(PeriodicOperation): |
There was a problem hiding this comment.
moved the periodic tasks into their own classes, which can be used/tested independently of the monitor thread
| PeriodicOperation("collect_and_send_events", self.collect_and_send_events, self.EVENT_COLLECTION_PERIOD), | ||
| PeriodicOperation("send_telemetry_heartbeat", self.send_telemetry_heartbeat, self.TELEMETRY_HEARTBEAT_PERIOD), | ||
| PeriodicOperation("poll_telemetry_metrics usage", self.poll_telemetry_metrics, self.CGROUP_TELEMETRY_POLLING_PERIOD), | ||
| ReportNetworkErrorsOperation(), |
There was a problem hiding this comment.
I left these 3 operations in the monitor class because I still need to unravel their dependency on the protocol
Codecov Report
@@ Coverage Diff @@
## develop #1908 +/- ##
===========================================
+ Coverage 69.38% 69.49% +0.10%
===========================================
Files 85 85
Lines 11813 11865 +52
Branches 1658 1666 +8
===========================================
+ Hits 8197 8246 +49
- Misses 3248 3249 +1
- Partials 368 370 +2
Continue to review full report at Codecov.
|
azurelinuxagent/common/cgroup.py
Outdated
|
|
||
| class MetricsCategory(object): | ||
| MEMORY_CATEGORY = "Memory" | ||
| PROCESS_CATEGORY = "Process" |
There was a problem hiding this comment.
Isn't a better name CPU? This metric contains CPU counters, right?
There was a problem hiding this comment.
I moved that code from elsewhere, but CPU does make more sense. We are not using the existing metric in kusto so I'll go ahead with the rename.
| @@ -1,3 +1,4 @@ | |||
| # -*- coding: utf-8 -*- | |||
There was a problem hiding this comment.
Ah, because of this?
├─27519 /usr/bin/python3 -u /usr/sbin/waagent -daemon
└─27547 python3 -u bin/WALinuxAgent-2.2.48.1-py2.7.egg -run-exthandlers
| @@ -130,6 +127,21 @@ def raise_exception(*_): | |||
| message = args[0] | |||
| self.assertIn("A TEST EXCEPTION", message) | |||
|
|
|||
| def test_get_processes_in_agent_cgroup_should_return_the_processes_within_the_agent_cgroup(self): | |||
There was a problem hiding this comment.
How is this test different from test_get_processes_in_cgroup_should_return_the_processes_within_the_cgroup in test_cgroupapi.py?
There was a problem hiding this comment.
code-wise they are the same, but test 2 different components (configurator vs api)
As part of collecting resource usage metrics we now report processes that should not be in the agent's cgroup.