Wazuh Dashboard: PoC - tests

[AlexRuiz7]

Description

Cypress tests are hosted into a different repository, and the CI workflow is configured to use pull requests' base branch as reference to checkout the tests' repository.

As we do not want to fork the tests' repository for the time being, a possible solution would be to patch the Cypress workflow to checkout the tests's repository at the version we want. given as an input (file, env variable...).

This issue aims to provide a proof of concept.

[AlexRuiz7]

Proof of concept: using GitHub vars

• Set up an environment variable at repository level
  Possible security concerns: variables render unmasked in your build outputs, but this isn't sensitive information.
• Read the variable in Actions
• Update the action ( 4d5f273, 923c132 )

   diff --git a/.github/workflows/cypress_workflow.yml b/.github/workflows/cypress_workflow.yml
   index 41b9d46aa3..53c39eb532 100644
   --- a/.github/workflows/cypress_workflow.yml
   +++ b/.github/workflows/cypress_workflow.yml
   @@ -13,6 +13,7 @@ env:
      OPENSEARCH_SNAPSHOT_CMD: 'node ../scripts/opensearch snapshot'
      SPEC: 'cypress/integration/core-opensearch-dashboards/opensearch-dashboards/**/*.js,'
      CYPRESS_ENV: 'env CYPRESS_VISBUILDER_ENABLED=true '
   +  OPENSEARCH_VERSION: ${{ vars.OPENSEARCH_VERSION }}
   
   jobs:
      cypress-tests:
  

[AlexRuiz7]

There are good and bad news. Good news is that the Cypress tests are launched correctly, hence confirming this PoC. Bad news is that most of the tests do fail (8/12).

(TODO)
At first, the changes to the action were not taking into account by, so I had to push them to the main branch as well. I expected the action of the base branch to be used, but this doesn't seem to be the case. I'll dive deeper in this matter.

In reference to the failed Cypress tests, they do fail by 3 reasons:

• Due to actual errors: something is broken and needs fixing.
• Due to our changes: tests need to be skipped until we find a better solution. In Cypress, this might be possible using the excludePatten option. In Jest however, this cannot be done as only test files or paths can be excluded, but not specific tests. See testMatch, testPathIgnorePattern and testRegex.
  We could either skip more tests than the necessary, or create a single test per file. In this case, we would need to manually divide all the tests, handle the errors and keep dividing them each time we receive new tests from upstream. None of the options are good, so we'll simply whitelist them and accept these tests to fail, watching no other tests than these fail.
• Due to timeout: the tests did not end in the expected time interval. Needs fixing.

These 3 cases will be treated in depth and one-by-one in a particular issue #34.

[AlexRuiz7]

I'll try to fix some of these tests and create a 2nd pull request to trigger a new run, in order to:

• check if the excludePatten option works.
• double check this PoC works.

[AlexRuiz7]

The PoC is working as stated in #33, and implemented in #34.