[Resolver] Fix flaky test #80576
Merged
[Resolver] Fix flaky test #80576
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Under stress, this test would fail to finish in 5 seconds. With this new implementation it has better performance.
oatkiller
commented
Oct 14, 2020
| /** | ||
| * An `enzyme` supported CSS selector for process node elements. | ||
| */ | ||
| public static nodeElementSelector({ |
There was a problem hiding this comment.
This was a module level function before. Adding it as a static field allows simulator tests to access it. This way they can use find on returned enzyme wrappers to find graph node elements.
| /** | ||
| * Given a `role`, return DOM nodes that have it. Use this to assert that ARIA roles are present as expected. | ||
| */ | ||
| public domNodesWithRole(role: string): ReactWrapper { |
There was a problem hiding this comment.
In order to test our ARIA compliance we need to select nodes by their role attribute.
There was a problem hiding this comment.
this will be useful - thanks.
oatkiller
added
release_note:skip
dplumlee
approved these changes
Oct 14, 2020
bkimmel
approved these changes
Oct 14, 2020
💚 Build SucceededMetrics [docs]
To update your PR or re-run it, just comment with: |
oatkiller
pushed a commit
to oatkiller/kibana
that referenced
this pull request
Oct 15, 2020
Under stress, this test would fail to finish in 5 seconds. With this new implementation it has better performance.
gmmorris
added a commit
to gmmorris/kibana
that referenced
this pull request
Oct 15, 2020
* master: (102 commits) [Resolver] Fix flaky test (elastic#80576) Update Security Solution Bug Report Template (elastic#80668) [Observability] Kibana home page Observability link pointing to `/landing` (elastic#80636) [APM] Update User Experience app callout code to reflect new name (elastic#80641) [APM] Add missing ML privileges (elastic#80553) [DOCS] Adds intro line to the ML plugin readme file (elastic#80631) [ML] Functional tests - fix and re-enable validation API tests (elastic#80617) remove non-existing dependency from uptime plugin (elastic#80623) [ML] Fix job selection flyout overflow (elastic#80621) Move dashboard code in codeowner files to canvas team (elastic#80345) [Security Solution][Detections] Update signals template if outdated and rollover indices (elastic#80019) Sort service list by TPM if health is not shown (elastic#80447) Add in cluster version for sec telemetry sender. (elastic#80545) [Usage Collection] Usage collection add saved objects client to collector fetch context (elastic#80554) Change tag from experimental to beta (elastic#80443) [Metrics UI] Inventory view cleanup (elastic#79881) [Security Solutions][Detection Engine] Critical bug where value lists were not operational (elastic#80368) [Security Solution] Fix networkTopNFlow search strategy response (elastic#80362) [build] Retry docker pull (elastic#80432) add template for Security Solution bugs (elastic#80574) ...
oatkiller
pushed a commit
that referenced
this pull request
Oct 15, 2020
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Under stress, this test would fail to finish in 5 seconds. With this new
implementation it has better performance.
I also exposed a new static method on the
Simulatorclass that creates CSS selectors for Resolver nodes and a new public method that selects DOM nodes via theroleattribute.Context
@XavierM noticed that this test occasionally failed on his local development environment. I was able to reproduce the issue by running the program
stress:I verified that my CPU was under intense load using
htop. Here is a partial screenshot. The bar charts 1 through 12 represent the CPU load of each of my computer's cores. "Load average" indicates how many of my cores my machine would have to have in order to complete all scheduled tasks concurrently (or something like that.)I then ran the
resolverjest test suite via the CLI and saw that the test in question failed:Depending on your environment, you may need to use a higher number in the command:
stress -c 32. Originally32worked for me, but at a later time I needed to use 65. It will depend on the other programs running on your machine.Once I reliably reproduced the test failure, I switched out the implementation. My assumption was that the call to
closestwas not very performant. The original implementation worked by searching a DOM tree for 3 nodes using a CSS selector and then searching those node's ancestors using another CSS selector. The new implementation searched for the ancestors first, then their descendants.I re-ran the tests with the same load. The test in question now passed.
The pattern of using
stressto include CPU load andhtopto confirm that the load is high was useful in reproducing this flaky performance related issue.Checklist
Delete any items that are not applicable to this PR.
For maintainers