Dress4Win has end-to-end tests covering 100% of their endpoints. They want to ensure that the move to the cloud does not introduce any new bugs. Which additional testing methods should the developers employ to prevent an outage?
A.
They should enable Google Stackdriver Debugger on the application code to show errors in the code.
B.
They should add additional unit tests and production scale load tests on their cloud staging environment.
C.
They should run the end-to-end tests in the cloud staging environment to determine if the code is working as intended.
D.
They should add canary tests so developers can measure how much of an impact the new release causes to latency.
Going with B. "Final answer" lol. C is a good answer except that they're asking for an additional testing method. Since they're already testing endpoints specifically, you'd literally be running the exact same test after migration. That said, for B, I'm still at a loss of why we'd need to do additional unit testing--best explanation is that some of the applications will have needed to be retooled for PaaS offerings if they're doing more than a lift-and-shift, thereby actually changing the underlying code; but the production-level load testing is like the most GCP thing you can do here
I would go with B. "Additional" seems to be the keyword here so adding unit tests and production scale load tests to the ones they already have makes more sense.
Question mention that "end-to-end tests covering 100% of their endpoints", "ensure that the move to the cloud does not introduce any new bugs", and "additional testing methods should the *developers* employ to *prevent an outage*"
A - Not Correct. Developer can debug the problem, but cannot *prevent* the outage.
B - Correct. Developers are responsible for writing unit tests. They already have end-to-end tests for *endpoints* but nothing mentioned about the unit tests. Cloud will auto-scale but you need to define your auto-scaling configuration (desired count, max count etc) and production scale load test will help you to configure the auto-scaling policies
C - Not Correct. They already have end-to-end test. Running it on staging environment will not prevent an outage
D - Not Correct. Answers says "an impact the new release causes to latency" but question ask for preventing an outage and so this one is ruled out
D
they already has 100% covering end-to-end testing, that means unit tests was finished before this
so, additional test remains canary in production environment
B is testing scale, C is testing the code is working as intended in the cloud. the question is for bugs, not for scale. It should be C, but I understand why everyone goes for B. Neurodivergents will go for C.
Note the ask "prevent an outage".
One of the way to test outage scenarios is through load testing. Option B covers this where as option C only covers checking the intended behaviour.
Testing pipeline: Unit ---- Integration --- end-to-end
If choose B, add additional unit tests. It should be continue to do Integrationt test and end-to-end test.
But it just has a unit test, then go to produciton scale load test.
C should be better, becasue it has finished end-to-end tests before, so for move to the cloud, it should be test the end-to-end in cloud preproduction enviroment to check whethere it's also working fine on cloud.
https://cloud.google.com/architecture/building-production-ready-data-pipelines-using-dataflow-developing-and-testing#end-to-end_tests
It should be B as for all those going with C if you do all staging you will still leave out the performance test that scales the application which is covered in B that means even if the application works well but will not scale properly will lead to an outage, which we are asked to prevent.
Question asks about "ADDITIONAL testing methods".
B adds production scale load tests.
C should also be executed in a new cloud env but this question dosn't ask for this.
In real life serious projects, B and C are mandatory (end to end and perf tests).
Should be B.
Should be B
Reasoning: They already had end-to-end contract tests coverage for all their service(s) endpoints. So, additionally they should add unit test coverage and perform prod load tests in staging environment which will help find out performance related issues before deploying it to production.
A voting comment increases the vote count for the chosen answer by one.
Upvoting a comment with a selected answer will also increase the vote count towards that answer by one.
So if you see a comment that you already agree with, you can upvote it instead of posting a new comment.
examtaker11
Highly Voted 4 years, 8 months agoJphix
3 years, 9 months agomesodan
2 years, 8 months agoSmart
4 years, 8 months agoFAB1010
Highly Voted 4 years, 3 months agobigzero
Most Recent 5 months, 2 weeks agoJconnor
11 months agoRitwickKumar
2 years, 2 months ago[Removed]
2 years, 7 months agojoe2211
2 years, 11 months agorottzy
3 years agoamxexam
3 years, 1 month agovictory108
3 years, 3 months agoMamthaSJ
3 years, 3 months agoAusias18
3 years, 7 months agolynx256
3 years, 7 months agoybe_gcp_cert
3 years, 7 months agoguid1984
3 years, 8 months agoRightsaidfred
3 years, 8 months agobnlcnd
3 years, 9 months ago