commit | a40fcbe45be1224e300d005d010e501a02a542cf | [log] [tgz] |
---|---|---|
author | Googler <lberki@google.com> | Mon Apr 22 01:53:26 2024 -0700 |
committer | Copybara-Service <copybara-worker@google.com> | Mon Apr 22 01:55:15 2024 -0700 |
tree | d7b62c4c2fb55f8e4240ef82cc0f003f3fbc8e17 | |
parent | ba8ccaf48a911f7bc9f2c18ab8f2fdb71d5c64ac [diff] |
Refactor event handling related to metrics collection and cleanup after commands. The end goal is to be able to report Skyframe statistics for commands that touch Skyframe but don't do builds. To this end, MetricsCollector had to be rigged up such that it reports Skyframe statistics in that case. Previously, that relied on Blaze emiting SkyframeGraphStatsEvent in BuildTool.stopRequest(), but "query" doesn't call that. That event is (confusingly) also emitted in BlazeRuntime.afterCommand(), but that was too late to be useful because that's after MetricsCollector has already reported the metrics. The new approach is as follows: 1. SkyframeExecutor.notifyCommandComplete() is always called exactly once, before BuildPrecompleteEvent. This is necessary because this is when SkyframeGraphStatsEvent is reported, which is where the Skyframe metrics data comes from. 2. BuildPrecompleteEvent is renamed to CommandPrecompleteEvent and is called for every command. As before, MetricsCollector reports metrics when this event is fired. 3. Metrics reporting pipelines are closed off in BlazeModule.afterCommand(). This required a tricky refactoring of some tests: * BlazeRuntime.afterCommand() is now called in integration tests so that the metrics reporting can happen. This is useful in any case because this way, what happens in tests is closer to what happens in production (we would ideally call BlazeCommandDispatcher but that's an even larger refactoring) * Some Google-internal tests relied on BlazeRuntime keeping state despite --nokeep_state_after_build. I added a special flag for them. Previously, this worked because BlazeRuntime.afterCommand() was not called and thus BlazeModule.afterCommand() was not called, either. * Some explicit calls to BlazeModule.afterCommand() were removed from tests because we now always call it. RELNOTES: None. PiperOrigin-RevId: 626956939 Change-Id: Ia15baab8701302c44ed5ea5f1156f8f57edc95f4
{Fast, Correct} - Choose two
Build and test software of any size, quickly and reliably.
Speed up your builds and tests: Bazel rebuilds only what is necessary. With advanced local and distributed caching, optimized dependency analysis and parallel execution, you get fast and incremental builds.
One tool, multiple languages: Build and test Java, C++, Android, iOS, Go, and a wide variety of other language platforms. Bazel runs on Windows, macOS, and Linux.
Scalable: Bazel helps you scale your organization, codebase, and continuous integration solution. It handles codebases of any size, in multiple repositories or a huge monorepo.
Extensible to your needs: Easily add support for new languages and platforms with Bazel's familiar extension language. Share and re-use language rules written by the growing Bazel community.
Follow our tutorials:
To report a security issue, please email security@bazel.build with a description of the issue, the steps you took to create the issue, affected versions, and, if known, mitigations for the issue. Our vulnerability management team will respond within 3 working days of your email. If the issue is confirmed as a vulnerability, we will open a Security Advisory. This project follows a 90 day disclosure timeline.
See CONTRIBUTING.md