HEAD
every four hours.When you click on a pipeline you can see the last few builds of this pipeline. Clicking on a build then gives you access to the details of the build. For example, the below image shows a failed build step on Ubuntu 16.04.
One can see which tests failed by clicking on the Test section. In the below example, the //src/test/shell/bazel:external_path_test
was flaky as it failed in 1 out of 5 runs.
You can view the failed test attempt's test.log
file in the Artifacts tab.
Bazel accepts contributions via pull requests. Contributions by members of the bazelbuild organisation as well as members of individual repositories (i.e. rule maintainers) are whitelisted automatically and will immediately be built and tested on Buildkite.
An external contribution, however, first needs to be verified by a project member and therefore will display a pending status named Verify Pull Request.
A member can verify a pull request by clicking on Details, followed by Verify Pull Request.
Please vet external contributions carefully as they can execute arbitrary code on our CI machines
After a pull request has been built and tested, the results will be displayed as a status message on the pull request. A detailed view is available when clicking on the corresponding Details link. Click here for an example.
You can preview the effect of an unmerged commit on downstream projects. See Testing Local Changes With All Downstream Projects.
Bazelisk + Incompatible flags pipeline runs bazelisk --migrate
on all downstream projects and reports a summary of all incompatible flags and migrations statuses of downstream projects.
The pipeline runs every night against the latest Bazel release. You can also schedule manual builds and set USE_BAZEL_VERSION
to run against a specific release or a release candidate (e.g. USE_BAZEL_VERSION=0.29rc3
to test against RC3 of release 0.29)
Bazel downstream projects is red? Use culprit finder to find out which bazel commit broke it!
First you should check if the project is green with the latest Bazel release. If not, probably it's their commits that broke the CI.
If a project is green with release Bazel but red with Bazel nightly, it means some Bazel commit broke it, then culprit finder can help!
Create “New Build” in the Culprit Finder project with the following environment variable:
,
. You can set this to bisect for multiple tasks in one build. It will be ignored if TASK_NAME is set.true
to run bazel clean --expunge
before each build, this will help reduce flakiness)eg.
PROJECT_NAME=rules_go PLATFORM_NAME=ubuntu2004 GOOD_BAZEL_COMMIT=b6ea3b6caa7f379778e74da33d1bd0ff6477f963 BAD_BAZEL_COMMIT=91eb3d207714af0ab1e5812252a0f10f40d6e4a8
Note: Bazel commit can only be set to commits after 63453bdbc6b05bd201375ee9e25b35010ae88aab, Culprit Finder needs to download Bazel at specific commit, but we didn't prebuild Bazel binaries before this commit.
Bazel Auto Sheriff is the pipeline to monitor Bazel CI build status and identify reasons for breakages.
Based on a project's build result in main build (with Bazel@Release) and downstream build (with Bazel@HEAD), the Bazel Auto Sheriff does analyzing by the following principles:
Main Build: PASSED, Downstream build: PASSED
Everything is fine.
Main Build: FAILED, Downstream build: PASSED
Retry the failed jobs to check if they are flaky
Main Build: PASSED, Downstream build: FAILED
Retry the failed downstream jobs to check if they are flaky
Main Build: FAILED, Downstream build: FAILED
Rebuild the project at last green commit
After the analysis, the pipeline will give a summary of four kinds of breakages:
You can check the analysis log for more details.
Each pipeline is configured via a Yaml file. This file either lives in $PROJECT_DIR/.bazelci/presubmit.yml
(for presubmits) or in an arbitrary location whose path or URL is passed to the CI script (as configured in the Buildkite settings of the respective pipeline). Projects should store the postsubmit configuration in their own repository, but we keep some configurations for downstream projects in https://github.com/bazelbuild/continuous-integration/tree/master/buildkite/pipelines.
The most important piece of the configuration file is the tasks
dictionary. Each task has a unique key, a platform and usually some build and/or test targets:
--- tasks: ubuntu_build_only: platform: ubuntu2004 build_targets: - "..." windows: platform: windows build_targets: - "..." test_targets: - "..."
If there is exactly one task per platform, you can omit the platform
field and use its value as task ID instead. The following code snippet is equivalent to the previous one:
--- tasks: ubuntu2004: build_targets: - "..." windows: build_targets: - "..." test_targets: - "..."
You can set environment variables for each individual task via the environment
field:
--- tasks: ubuntu1804: environment: CC: clang build_targets: - "..."
The presubmit configuration allows you to specify a list of shell commands that are executed at the beginning of every job. Simply add the batch_commands
(Windows) or shell_commands
field (all other platforms).
You can even run executable targets via the run_targets
field. The following example demonstrates all of these features:
--- tasks: ubuntu1804: shell_commands: - rm -f obsolete_file run_targets: - "//whatever" build_targets: - "..." windows: batch_commands: - powershell -Command "..." build_targets: - "..."
The build_flags
and test_flags
fields contain lists of flags that should be used when building or testing (respectively):
--- tasks: ubuntu1804: build_flags: - "--define=ij_product=clion-latest" build_targets: - "..." test_flags: - "--define=ij_product=clion-latest" test_targets: - ":clwb_tests"
Each task may have an optional display name that can include Emojis. This feature is especially useful if you have several tasks that run on the same platform, but use different Bazel binaries. Simply set the name
field:
--- tasks: windows: name: "some :emoji:" build_targets: - "..."
You can use kythe_ubuntu2004
platform along with some index_*
fields to create a task that generate semantic information of your code with Kythe.
The index_targets
field contains list of targets that should be indexed.
The index_targets_query
field contains a query string that is used to generate additional index targets by command bazel query ${index_targets_query}
. The returned targets will be merged into index_targets
.
The index_flags
field contains list of build flags that should be used when indexing.
The index_upload_policy
field is used to set policy for uploading generated files. Available values are:
Always
: Always upload generated files even build failed.IfBuildSuccess
: Default value. Only upload generated files if build succeed and raise error when build failed.Never
: Never upload index files and raise error when build failed.If index_upload_gcs
is True
, the generated files will be uploaded to Google Cloud Storage.
--- tasks: kythe_ubuntu2004: index_targets: - "..." index_targets_query: "kind(\"java_(binary|import|library|plugin|test|proto_library) rule\", ...)" index_flags: - "--define=kythe_corpus=github.com/bazelbuild/bazel" index_upload_policy: "IfBuildSuccess" index_upload_gcs: True
Most existing configuration use the legacy format with a “platforms” dictionary:
--- platforms: ubuntu1804: build_targets: - "..." test_targets: - "..."
The new format expects a “tasks” dictionary instead:
--- tasks: arbitrary_id: platform: ubuntu1804 build_targets: - "..." test_targets: - "..."
In this case we can omit the platform
field since there is a 1:1 mapping between tasks and platforms. Consequently, the format looks almost identical to the old one:
--- tasks: ubuntu1804: build_targets: - "..." test_targets: - "..."
The CI script still supports the legacy format, too.
The CI uses Bazelisk to support older versions of Bazel, too. You can specify a Bazel version for each pipeline (or even for individual platforms) in the pipeline Yaml configuration:
--- bazel: 0.20.0 tasks: windows: build_targets: - "..." macos: build_targets: - "..." ubuntu1804: bazel: 0.18.0 build_targets: - "..." [...]
In this example the jobs on Windows and MacOS would use 0.20.0, whereas the job on Ubuntu would run 0.18.0.
CI supports several magic version values such as latest
, last_green
and last_downstream_green
. Please see the Bazelisk documentation for more details.
Similar to the aforementioned downstream pipeline you can configure individual pipelines to run with bazelisk --migrate
. As a result, the pipeline runs your targets with all incompatible flags that will be flipped in the next major Bazel release and prints detailed information about which flags need to be migrated.
You can enable this feature by adding the following code to the top of the pipeline steps in Buildkite at https://buildkite.com/bazel/YOUR_PIPELINE_SLUG/settings, not in the pipeline configuration yaml file:
--- env: USE_BAZELISK_MIGRATE: true
If you want your pipeline to fail if at least one flag needs migration, you need to add this code instead:
--- env: USE_BAZELISK_MIGRATE: FAIL
If you want to enable this feature for a single build, but not for the entire pipeline, you can follow these steps instead:
USE_BAZELISK_MIGRATE=FAIL
into the “Environment Variables” text field.We upgrade the CI machines to the latest version of Xcode shortly after it is released and this version will then be used as the default Xcode version. If required, you can specify a fixed Xcode version to test against in your pipeline config.
Warning: We might have to run jobs that specify an explicit Xcode version on separate, slower machines, so we really advise you to not use this feature unless necessary.
The general policy is to not specify a fixed Xcode version number, so that we can update the default version more easily and don't have to update every single CI configuration file out there.
However, if you know that you need to test against multiple versions of Xcode or that newer versions frequently break you, you can use this feature.
tasks: # Test against the latest released Xcode version. macos: build_targets: - "..." # Ensure that we're still supporting Xcode 10.1. macos_xcode_10_1: platform: macos xcode_version: "10.1" build_targets: - "..."
Take care to quote the version number, otherwise YAML will interpret it as a floating point number.
For each pipeline you can enable Buildifier to check all WORKSPACE, BUILD, BUILD.bazel and .bzl files for lint warnings and formatting violations. Simply add the following code to the top of the particular pipeline configuration:
--- buildifier: latest [...]
As a consequence, every future build for this pipeline will contain an additional “Buildifier” step that runs the latest version of Buildifier both in “lint” and “check” mode. Alternatively you can specify a particular Buildifier version such as “0.20.0”.
There is also a more advanced syntax that allows you to specify which warnings should be checked in lint mode:
--- buildifier: version: latest warnings: "positional-args,duplicated-name" [...]
Some projects may contain one or more WORKSPACE
files in subdirectories, in addition to their top-level WORKSPACE
file. All of these workspaces can be tested in a single pipeline by using the working_directory
task property. Consider the configuration for a project that contains a second WORKSPACE
file in the examples_dir/
directory:
--- tasks: production_code: name: "My Project" platform: ubuntu1804 test_targets: - //... examples: name: Examples platform: ubuntu1804 working_directory: examples_dir test_targets: - //...
You can set the top-level validate_config
option to ensure that changes to pipeline configuration files in the .bazelci
directory will be validated. With this option, every build for a commit that touches a configuration file will contain an additional validation step for each modified configuration file.
Example usage:
--- validate_config: 1 tasks: macos: build_targets: - "..."
Bazel's JSON Profile is a useful tool to investigate the performance of Bazel. You can configure your pipeline to export these JSON profiles on builds and tests using the include_json_profile
option.
Example usage:
--- tasks: ubuntu2004: include_json_profile: - build - test build_targets: - "..." test_targets: - "..."
When include_json_profile
is specified with build
, the builds will be carried out with the extra JSON profile flags. Similarly for test
. Other values will be ignored.
The exported JSON profiles are available as artifacts after each run.
Q: I added or changed an external repository and now my test is failing on Bazel CI only with errors like this:
WARNING: Download from https://github.com/bazelbuild/java_tools/releases/download/javac11-v11.0/java_tools_javac11_linux-v11.0.zip failed: class java.net.ConnectException Operation not permitted (connect failed) ERROR: An error occurred during the fetch of repository 'remote_java_tools_linux_beta': java.io.IOException: Error downloading [https://github.com/bazelbuild/java_tools/releases/download/javac11-v11.0/java_tools_javac11_linux-v11.0.zip] to /private/var/tmp/_bazel_buildkite/c3a616e1648c5e14a8ab09d0d59696c2/sandbox/darwin-sandbox/3279/execroot/io_bazel/_tmp/58d272c7f3dd803b2bcb2fc7be47d391/root/fb8b458bcc92813a6fcf57a0dbe6e8bd/external/remote_java_tools_linux_beta/java_tools_javac11_linux-v11.0.zip: Operation not permitted (connect failed)
A: We run most tests on CI without network access and instead inject the external repositories from the outside. This saves a lot of network traffic and I/O (because the Bazel integration tests don't have to extract the repository archives again and again).
In the code review of this PR, philwo@ explained how to fix test failures like this: https://github.com/bazelbuild/bazel/pull/11436.