wolfseb

Hi @emakeev, could you please take another look at this if you get the chance to see whether the updates have addressed your comments? Thanks!

We did a lot of pinning of the dependencies in requirements.in as well, which you can find in the files changed section of #13482. We had to pin some dependencies...

Update: we got all magma services running by reinstalling the following Python packages: * prometheus-client * h2 * jinja2 * eventlet * redis * ovs * envoy * click via...

Integ tests have been run for the new VM. The following tests fail: Precommit: None Extended: None Nonsanity: - test_activate_deactivate_multiple_dedicated.py - test_sctp_shutdown_while_mme_is_stopped.py - test_3495_timer_for_dedicated_bearer_with_mme_restart.py - test_no_attach_complete_with_mme_restart.py

With the changes in #13659, the integ test `test_attach_detach.py` works against the containerized AGW also after the python dependency upgrade.

Do I understand it correctly that the new logic retries starting the containers when one of them is briefly unhealthy after startup? (Essentially it would just restart the waitForHealthyPods container...

Looking through the logs of the failing runs, the first of the three failures (-> flaky retry) is always due to `Assertion Error: 14 != 28`

After some testing on the github runner and locally, we found the following: * If the time between initiating the mme restart and the `get_response` is longer than 30s, the...