优化构建与测试性能

Welcome! In large monorepos, build and test times can become a major bottleneck. Slow feedback loops hurt developer productivity and can delay releases.

This lesson will show you how to significantly speed up these processes by applying advanced optimization techniques.

Why do monorepos often struggle with performance?

• Scale: A single repository can contain hundreds of projects.
• Interdependencies: A small change in one project might affect many others.
• Traditional CI: Often rebuilds and retests everything on every commit, which is highly inefficient.

Imagine your computer remembering the exact output of a task it just did. That's build caching!

• It stores the results (artifacts) of previous build or test commands.
• If the inputs (code, configurations) haven't changed, it reuses the stored output.
• This avoids redundant work and saves a lot of time.

Monorepo tools like Nx and Turborepo implement local caching automatically. When you run a command, they check a local cache first.

If a cached result exists for the exact inputs, it's retrieved instantly instead of re-running the task.

Local caching is great for individual developers, but teams need more. Remote caching shares these cached build artifacts across everyone!

• If a project was built or tested by a teammate or CI, you can download the cached result.
• This means fewer builds from scratch for the entire team and faster CI/CD pipelines.

Why rebuild your entire frontend application if you only changed a backend API? This is where incremental builds come in.

Also known as "affected" commands, they identify and process only the projects that have been impacted by your recent code changes.

Monorepo tools create a dependency graph – a map of how all your projects connect.

When you make a change, the tool uses this graph to intelligently determine:

• Which projects were directly modified.
• Which other projects depend on the modified ones.

Only these "affected" projects are then built or tested.

Even with caching and incremental builds, some tasks are just computationally intensive. Remote execution helps by distributing parts of a build or test command across multiple machines.

This turns a single, long-running task into many smaller, parallel tasks that complete much faster.

Remote execution significantly boosts performance by:

• Parallelization: Running many tasks simultaneously.
• Resource Scaling: Leveraging cloud resources or dedicated build farms.
• Faster Feedback: Getting build and test results back much quicker.

It's especially powerful for large test suites and complex compilation steps.

Which technique focuses on reusing the outputs of previous tasks to avoid redundant work, assuming the inputs haven't changed?

You've learned powerful strategies to optimize build and test performance in monorepos:

• Build Caching: Store and reuse results of past tasks, both locally and remotely.
• Incremental Builds: Use dependency graphs to only process projects affected by changes.
• Remote Execution: Distribute heavy workloads across multiple machines for parallel processing.

Mastering these techniques is essential for maintaining developer productivity in any growing monorepo.