-
Notifications
You must be signed in to change notification settings - Fork 27
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[CSM O11y test] Ping GMP endpoint during test for debugging purpose #33
Conversation
- Ping GMP endpoint in client and server pod curl localhost:9464/metrics during test - This will provide insight into what GMP gets from the OTel plugin before sending metrics to Cloud Monitoring
- But this doesn't work yet. Can't quite figure out what's the exact request URL to use. Tried quite a few combinations and none of them works.
Do you know what the request URL should I be using? I tried quite a few combinations but none of them works
If I supply I just realized I am running the |
- `enable_csm_observability` is now a constructor parameter for the client runner and the server runner, instead of a `run()` parameter. - If `enable_csm_observability`, we want to do a port forwarding for the `pod_monitoring_port = 9464`, such that we can issue a GET request against it in the CSM test - Added `monitoring_port` and `monitoring_host` as class variable to the client runner and server runner. - Override `_xds_test_server_for_pod` in the `GammaServerRunner` class - Minimize the places where we need to hardcode the port `9464`
- Added a PrometheusLogger class to write the prometheus endpoint log to a separate log file related to logs_subdir - monitoring_port is now stored in XdsTestClient and XdsTestServer rather than the runner - use rpc_host instead of a separate monitoring_host - put the constant DEFAULT_MONITORING_PORT in the KubernetesBaseRunner - fixed requirements.txt and requirements.lock - removed a previous hack is_legit_time_series() - renamed function to ping_prometheus_endpoint instead of referring it as GMP
2bb2bd3
to
c71a050
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM with
- FIxing the log collection arg
- Running full csm test suite + lb test suite
@sergiitk Incorporated latest feedback. Using Want to do a final check and merge? Thanks. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It appears that a bunch grpc/core/master/linux/psm-csm test run failed with
Traceback (most recent call last):
File "/tmp/tmp.2ObQJ9ydNu/psm-interop/tests/app_net_ssa_test.py", line 111, in test_session_affinity_policy
test_client: _XdsTestClient = self.startTestClient(
File "/tmp/tmp.2ObQJ9ydNu/psm-interop/framework/xds_k8s_testcase.py", line 928, in startTestClient
return self._start_test_client(test_server.xds_uri, **kwargs)
File "/tmp/tmp.2ObQJ9ydNu/psm-interop/framework/xds_k8s_testcase.py", line 831, in _start_test_client
test_client = self.client_runner.run(
File "/tmp/tmp.2ObQJ9ydNu/psm-interop/framework/test_app/runners/k8s/k8s_xds_client_runner.py", line 188, in run
**self.deployment_args.as_dict(),
AttributeError: 'NoneType' object has no attribute 'as_dict'
The lb test suite failed with the same error. |
- Created a new test `app_net_csm_observability_test.py` that's largely a copy of `gamma/csm_observability_test.py` with some adjustments. - Refactor some flags to be passed to the `k8s_xds_server_runner`. - Use `ServerDeploymentArgs` everywhere This PR was based on top of #33, and a replacement of #72.
<pod IP>:9464/metrics
during the CSM Observability test