[Reporting] Better logging for waitForSelector failure #25762
Merged
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
tsullivan
added
WIP
tsullivan
changed the title
💔 Build Failed |
|
Edit: see #25762 (comment) |
💔 Build Failed |
💔 Build Failed |
💔 Build Failed |
|
Will log all the plain text found on the page, when a job has to retry because a selector wasn't found. The lines of text are joined into one log line. Example: |
💔 Build Failed |
| @@ -29,10 +29,6 @@ export default function ({ getService }) { | |||
| expect(usage.reporting.enabled).to.be(true); | |||
| }); | |||
|
|
|||
| // "Kibana could not load correctly", etc | ||
| this.logger.error(`waitForSelector ${selector} failed on ${this.page.url()}`); | ||
| const pageText = await this.evaluate({ | ||
| fn: () => document.querySelector('body')!.innerText, |
There was a problem hiding this comment.
You probably don't want the bang here:
fn: () => document.querySelector('body').innerText,
There was a problem hiding this comment.
The ! is the "Non-null assertion operator" for Typescript to not fail compilation for the reason that the return from document.querySelector('body') is possibly null. It's not as safe as a guard, though. It's possible for querySelector to return null if an element isn't found, but for body I don't see it ever happening.
There was a problem hiding this comment.
TIL! Does the TS compiler remove this when built? Reason I ask is because this function gets stringified and sent over WebSockets over to the browser... so if it doesn't it'll have issues running
joelgriffith
suggested changes
Nov 19, 2018
💚 Build Succeeded |
💚 Build Succeeded |
joelgriffith
approved these changes
Nov 27, 2018
Yes, it doesn't have any effect on the built JS, it's just a compile-time operator. |
tsullivan
added a commit
to tsullivan/kibana
that referenced
this pull request
Nov 27, 2018
* [Reporting] Better logging for waitForSelector failure * break waitForSelector * experimental changes * cleanup/consistency * fix some test report title strings * test disable chromium * roll back test code * take out non-working phantom changes * roll back disable chromium test * allow logger to use curried tags * temporarily re-do report failure-causing change for test * replace newline with escaped for single log line * undo test change * remove obsolete test
tsullivan
added a commit
that referenced
this pull request
Nov 27, 2018
* [Reporting] Better logging for waitForSelector failure * break waitForSelector * experimental changes * cleanup/consistency * fix some test report title strings * test disable chromium * roll back test code * take out non-working phantom changes * roll back disable chromium test * allow logger to use curried tags * temporarily re-do report failure-causing change for test * replace newline with escaped for single log line * undo test change * remove obsolete test
sebelga
added a commit
that referenced
this pull request
Nov 28, 2018
* [APM] Fix horizontal scrollbar being visible in windows 8.1 (#25988) * [APM] Changed 'Response Time' to 'Duration' in transactions screens (#25990) * translate InfraOps visualization component (Part 3) (#25213) * translate InfraOps visualization component (Part 3 - part of folder components) * update translation of Infra Ops vizualization component (Part 3) * update translation of Infra Ops vizualization component (Part 3) * change some ids and add pluralization * update Infra Ops Part 3 - change some ids, change some intl.formatMessage() to <FormattedMessage> and directly wrap some classes by injectI18n() * update Infra-III - add static to displayName * [i18n] Translate Agg_types(part_3) (#26118) * Translate agg_types - metrics * Fix issues * [ML] Aggregate anomalies table data using configured Kibana timezone (#26192) * [ML] Aggregate anomalies table data using configured Kibana timezone * [ML] Move dataFormatTz prop out of controller scope * [ML] Fix alignment of filter icons in anomalies table (#26253) * [ML] Fix alignment of filter icons in anomalies table * [ML] Teak y position of icons in expanded row of table * translate sample data (#26069) translate sample data * [ML] Wrap controller initialization in assertions. (#26265) - The controller tests introduced in #25382 had a flaw: If a controller initialization would fail and throw an error, that test suite wouldn't be able to clean up any stubs. So tests using the same stubs would report and error because the stubs couldn't be wrapped again. - This PR wraps every controller initialization inside an assertion and catches those errors properly as part of the test. * [APM] fixes #20145 by displaying span.context.http.url in the span details flyout (#26238) * Fix spaces license check (#26270) ## Summary Allows the public spaces API to work with a gold license Resolves #26271 * Job Info button in Reporting Listing (#25421) * Job Info button in Reporting Listing * use lodash directly * start of flyout use * description list in flyout * capitalize * undefined guard * expire info on close * add jest test * better at error handling + messaging * Add description for vis types (#26243) * [migrations] Throw error if reindex task fails (#26062) Signed-off-by: Tyler Smalley <[email protected]> * [Reporting] Better logging for waitForSelector failure (#25762) * [Reporting] Better logging for waitForSelector failure * break waitForSelector * experimental changes * cleanup/consistency * fix some test report title strings * test disable chromium * roll back test code * take out non-working phantom changes * roll back disable chromium test * allow logger to use curried tags * temporarily re-do report failure-causing change for test * replace newline with escaped for single log line * undo test change * remove obsolete test * [kbn/pm] allow packages to define extra paths to clean (#26132) I noticed some discussion about how kbn clean should probably clear out the `.eslintcache` file, since it doesn't handle changes in related modules (for things like the import plugin) and it would be nice if `yarn kbn clean` took care of the issue. I figured it's not a bad idea, but adding `.eslintcache` directly to `@kbn/pm` felt wrong, so instead I've added another config options that can go in the package.json file, `clean.extraPatterns`. This array of patterns is passed into `del()` so that it can include things like negation. As the name suggests this doesn't override the initial paths, just adds some extras that will be checked and cleared when `yarn kbn clean` is run. * [config] fix logging.useUTC deprecation unset (#26053) * Extend precommit hook script to support git GUI apps (#25883) * feat(NA): extend support from precommit hook to git GUI apps. * docs(NA): more descriptive error message. * [DOCS] Clarify monitoring dependencies (#26229) * apm: add ECS fields to index pattern (#26214) * support standard license (#26294) * [kbn-pm] update build * [eslint] use disallow license header rule (#26309) Fixes #26295 There are several places where we have accidentally added new license headers with linters but failed to remove old license headers manually. This prevents that by applying the an inverted version of the license headers rule that removed invalid license headers when files are moved. * Bump node to 8.14.0 (#26313) Signed-off-by: Tyler Smalley <[email protected]> * Watch optimizer cache invalidation (#24172) * chore(NA): cherry pick work from spencer on impleting the cache invalidation system and merging it with current master. * feat(NA): add support for dlls bundle into the cache state invalidation system. * chore(NA): merge with master. * feat(NA): first working version for the watch cache. * feat(NA): added logger, correct cache delete and removed last todos. * feat(NA): remove some useless features for the time being. * refact(NA): just pass kibanaHapiServer.log function directly instead of an anonimous function that calls the kibanaHapiServer.log one. * refact(NA): move everything to async. * refact(NA): remove dll mentions. * chore(NA): removed types/mkdirp as we dont use mkdirp into typescript.
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.
Part of #25665
Summary
If a headless browser fails because the expected page couldn't be opened, we detect that by trying to find a CSS selector, for example:
.application. The error handling and messaging from the headless browser take over if that fails. That is pretty useless because we'll be told that there was a timeout looking for the selector when the real problem is the page failed to load.This change adds more logging:
Initially, this page changes the
applicationclass on the main HTML template, so we can check reporting failures in Jenkins to ensure they make more sense