Implemented HealthCheck Endpoint and improved Mailman resilience #3
Conversation
The /v1/moc/HealthCheck endpoint will perform checks for various external components that the service needs access to. Ex. SSH connection to the mailing list server, which is the primary source of flakiness at the moment. Additionally, refactored the Mailman code into a separate class, to facilitate testing and health checking. Finally, made failing to subscribe a user to a mailing list not be a blocking error.
| if errors: | ||
| raise ConnectionError(errors) | ||
|
|
||
| # Note(knikolla): Not entirely sure if closing before reading is fine. |
There was a problem hiding this comment.
is this comment old, or is it referring to something that exec_command is doing?
There was a problem hiding this comment.
That comment is old. I also realized that I didn't delete the old unused methods in the action, which I was keeping as a reference while splitting out the functionality to its separate class. Will delete both the comment and those now.
|
With regards to your comment about TODOs, it makes sense to me moving forward now that we have an established issue/task/sprint tracker. |
| ] | ||
| ) | ||
|
|
||
| def get(self, request, format=None): |
There was a problem hiding this comment.
request and format are unused.
format is a built-in so you might wanna call it something else if you do want to use that argument.
| from confspirator import groups as config_groups | ||
| from confspirator import fields as config_fields | ||
| from rest_framework.response import Response | ||
| from adjutant.api import utils |
|
|
||
| return Response({'errors': errors}) | ||
|
|
||
| def _get_action_defaults(self, action): |
| self.client.close() | ||
|
|
||
| def _execute(self, command): | ||
| stdin, stdout, stderr = self.client.exec_command(command) |
There was a problem hiding this comment.
nit: stdin is unused. _ can be used for throaway values
The /v1/moc/HealthCheck endpoint will perform checks for various
external components that the service needs access to. Ex. SSH
connection to the mailing list server, which is the primary
source of flakiness at the moment.
Additionally, refactored the Mailman code into a separate class, to
facilitate testing and health checking.
Finally, made failing to subscribe a user to a mailing list not
be a blocking error.