* [build] Move canvas runtime to core build, optionally skip step
- Moves the x-pack/build canvas runtime to src/dev/build. The current
workflow is an artifact of the separate oss and default builds, which
were removed in 7.10.
- Adds a new flag `node scripts/build --skip-canvas-shareable-runtime`.
Currently the runtime takes ~3 minutes to build, and as far as I can
tell is untested in pull requests. This should save us a few minutes
and CPU cycles, while preserving compatibility with upstream jobs.
* more flexible build args
* fix
* fix
* spacing
* trim spaces
* label
* Update .buildkite/scripts/build_kibana.sh
Co-authored-by: Spencer <email@spalger.com>
* remove preserve-symlinks
* fix ci stats plugin
Co-authored-by: Spencer <email@spalger.com>
* [kbn-performance-testing-dataset-extractor] fetch & save ES queries
* run apm extractor for all journeys
* fix array name
* cleanup ES query, update namings according review
* update naming
* [ci] Conditional storybook builds
This moves storybook builds to run conditionally, either if there's a
storybook change or the `ci:build-storybooks` label is set.
These builds don't seem to be providing test coverage, and may not be utilized
in scenarios such as automated backports or unrelated changes.
* test regex
* formatting
* Revert "test regex"
This reverts commit dcef465414.
* fix revert
* move to spot instance
* fix merge
* [cft] Exit 1 if purge error
This triggers a job failure if the purge-cloud-deployments script
encounters an error. I'm attempting to debug why an instance was stuck
online and hoping this will make future issues easier to spot.
* set exit code instead of exiting
* [kbn-performance-testing-dataset-extractor] update json structure, filter out static res on ci
* convert body to string
* update schema and ftr configs
* update extractor
* fix headers combining
* update json structure
* re-order fields for easy read
* [kbn/pm] rewrite to avoid needing a build process
* uncomment timing reporting
* throw in a few missing comments
* Update README.md
* remove extra SomeDevLog interface from ci-stats-core
* remove non-stdio logging from bazel_runner, improve output formatting
* use private fields instead of just ts private props
* promote args to a positional arg
* optionally require the ci-stats-reporter after each command
* allow opt-ing out of vscode config management
* reduce to a single import
* add bit of docs regarding weird imports and package deps of kbn/pm
* clean extraDirs from Kibana's package.json file too
* tweak logging of run-in-packages to use --quiet and not just CI=true
* remove unlazy-loader
* add readme for @kbn/yarn-lock-validator
* convert @kbn/some-dev-logs docs to mdx
* remove missing navigation id and fix id in dev-cli-runner docs
* fix title of some-dev-logs docs page
* typo
* bump ansible logging verbosity
* upgrade base os
* update base images
* add specificity to docker image glob
* limit memory
* bump base vm memory
* more cpu
* decrease verbosity
* less downloads
* disable shared symlinks
* more logging
* bump concurrency to test stability
* add concurrency group
* revert
* parallelism
* Revert "parallelism"
This reverts commit 4cc00f7144.
* add scalability setup for login and promotion_tracking_dashboard journeys, re-enable step on ci
* update query to get single ftr transaction
Co-authored-by: Kibana Machine <42973632+kibanamachine@users.noreply.github.com>
* [QA][Code Coverage] Add a link to discover pinned to the current build.
When personnel other than myself wish to visually verify code coverage,
this link will be available in the ci output.
* Drop single quotes.
* Add annotation per J Budz.
* Fixup text.
* Drop debug.
## Summary
Fixes issue with security solution cypress flaky test runner
After introducing [dynamic split for cypress tests in Security Solution](https://github.com/elastic/kibana/pull/125986), there was discovered an issue with flaky test runner, which is using parallelism in slightly different manner: https://buildkite.com/elastic/kibana-flaky-test-suite-runner/builds/697#_: tests were evenly split between N jobs, instead of repeating tests N times(as N jobs)
So, for flaky runner, I introducing a new ENV variable that would disable split for tests between parallel jobs and instead would run all test per each job.
Here is link to flaky runner build from this PR, which shows that all test runs per job
https://buildkite.com/elastic/kibana-flaky-test-suite-runner/builds/736
* [artifacts] Fallback to port 443 when testing cloud deployments
The provided endpoint for ES from cloud no longer contains a port. When
a port is not provided to the FTR, it defaults to 9220 causing a
connection timeout. This sets the fallback to 443 instead.
* use process.env instead of interpolation
Kinda sneaky...since we already mutate the jest portion of the
file system (target/kibana-coverage/jest)
by dumping "jest unit" & "jest integration" coverage
into the same "final" directory...go ahead an make "jest integration"
use the same ran file designator as "jest unit".
This saves me from having to add logic for this later on.
This test was added as a verification step for daily releasable
artifacts. We had originally planned on the release manager consuming
this context. The expected workflow has changed now to us building and
publishing the docker image instead, which already has coverage in the
Build and Deploy to Cloud step.
