Fix/increase nginx timeout

[kate-osborn]

Proposed changes

Problems:

1. The conformance tests consistently fail during setup because NKG times out waiting for the nginx process to start.
2. The conformance test makefile command runs the incorrect set of tests and times out before running all tests.

Solutions:

1. Increase the pid timeout (i.e. the time we wait for nginx to start) to 10s. This is sufficient in my local environment.
2. Update the list of supported features and exempt features and pass them to the conformance test suite, and increase the go test timeout to 25 minutes to allow the suite to complete. I also added the debug flag to increase the verbosity of the test output.

Testing: Ran the conformance tests locally. Confirmed that they will consistently pass setup and run to completion.

Checklist

• I have read the CONTRIBUTING doc
• I have added tests that prove my fix is effective or that my feature works
• I have checked that all unit tests pass after adding my changes
• I have updated necessary documentation
• I have rebased my branch onto main
• I will ensure my PR is targeting the main branch and pulling from my branch from my own fork

[pleshakov]

Increase the pid timeout (i.e. the time we wait for nginx to start) to 10s. This is sufficient in my local environment.

I wonder if the timeout also depends on the time it takes for Docker to download NGINX image. if that is the case, it might take more than 10s in different environments and perhaps pre-loading NGINX image can also help?

perhaps the big problem here is that NKG once it fails to reload, doesn't try to recover from it and reload again? (unless the config is re-generated for some reason)
if NKG supported retrying, it would be able to wait until NKG started and eventually fix the statuses. I think being able to fix that will be important not only for conformance tests but for production (since the timeout 10s might not be enough)

at the same time, approving this PR, since it fixes the issue for the conformance tests and the other changes also make sense 👍

[kate-osborn]

Increase the pid timeout (i.e. the time we wait for nginx to start) to 10s. This is sufficient in my local environment.

I wonder if the timeout also depends on the time it takes for Docker to download NGINX image. if that is the case, it might take more than 10s in different environments and perhaps pre-loading NGINX image can also help?

perhaps the big problem here is that NKG once it fails to reload, doesn't try to recover from it and reload again? (unless the config is re-generated for some reason) if NKG supported retrying, it would be able to wait until NKG started and eventually fix the statuses. I think being able to fix that will be important not only for conformance tests but for production (since the timeout 10s might not be enough)

at the same time, approving this PR, since it fixes the issue for the conformance tests and the other changes also make sense 👍

perhaps pre-loading NGINX image can also help

I think that's a smart idea. We should do that in our conformance test pipeline job @ciarams87 @lucacome

perhaps the big problem here is that NKG once it fails to reload, doesn't try to recover from it and reload again?

Yep, we need to handle this. Do we have a ticket for this yet? If not, I can create one.

[pleshakov]

@kate-osborn

Yep, we need to handle this. Do we have a ticket for this yet? If not, I can create one.

we have a related #664 but nothing about retrying reloads

[kate-osborn]

@kate-osborn

Yep, we need to handle this. Do we have a ticket for this yet? If not, I can create one.

we have a related #664 but nothing about retrying reloads

Updated the ticket to include retries