MEMO

pre-requisite

#

1. Extension Pack for Java (Microsoft)
2. Debugger for Java (usually included via the pack)
3. Also make sure VS Code is using your JDK (Java 17/21 etc):
4. Ctrl+Shift+P → Java: Configure Java Runtime
5. Ensure JAVA_HOME is set (System env) and matches what you want.

.vscode/tasks.json

#

{
  "version": "2.0.0",
  "tasks": [
    {
      "label": "boot:run (debug 5005)",
      "type": "shell",
      "command": "mvn -q -DskipTests spring-boot:run -Dspring-boot.run.jvmArguments=\"-agentlib:jdwp=transport=dt_socket,server=y,suspend=n,address=127.0.0.1:5005\"",
      "options": {
        "cwd": "${workspaceFolder}"
      },
      "problemMatcher": []
    },
    {
      "label": "boot:run (normal)",
      "type": "shell",
      "command": "mvn -q -DskipTests spring-boot:run",
      "options": {
        "cwd": "${workspaceFolder}"
      },
      "problemMatcher": []
    }
  ]
}

.vscode/launch.json

#

{
  "version": "0.2.0",
  "configurations": [
    {
      "name": "Attach to Spring Boot (5005)",
      "type": "java",
      "request": "attach",
      "hostName": "127.0.0.1",
      "port": 5005
    }
  ]
}

profile

#

mvn spring-boot:run ^
  -Dspring-boot.run.profiles=local ^
  -Dspring-boot.run.jvmArguments="-agentlib:jdwp=transport=dt_socket,server=y,suspend=n,address=127.0.0.1:5005"

x

#

With 16GB RAM + big Java/Maven repo, your goal is: reduce file watching/indexing, cap Java language server, and avoid scanning build outputs. Do these in this order.

1) Exclude build outputs everywhere (highest impact)

#

Open Settings (JSON) and add/merge:

{
  "files.watcherExclude": {
    "**/.git/**": true,
    "**/target/**": true,
    "**/build/**": true,
    "**/out/**": true,
    "**/.gradle/**": true,
    "**/.mvn/**": false
  },
  "search.exclude": {
    "**/target/**": true,
    "**/build/**": true,
    "**/out/**": true,
    "**/node_modules/**": true
  },
  "files.exclude": {
    "**/target/**": true,
    "**/build/**": true
  }
}

Why: on big repos, target/ churn is the #1 VS Code killer.

2) Cap Java Language Server memory (don’t let it eat your machine)

#

Add:

{
  "java.jdt.ls.vmargs": "-Xms256m -Xmx2048m -XX:+UseG1GC -XX:+UseStringDeduplication -Dsun.zip.disableMemoryMapping=true"
}

• -Xmx2048m is a good ceiling on 16GB (you still need RAM for Chrome, Maven builds, etc.).
• If you notice OOM in Java features, bump to 3072m, but don’t go higher unless you must.

When things get sluggish:

• Ctrl+Shift+P → Java: Clean Java Language Server Workspace (it often fixes “mysterious lag”).

3) Kill expensive UI features (small but steady wins)

#

{
  "editor.minimap.enabled": false,
  "editor.codeLens": false,
  "breadcrumbs.enabled": false,
  "git.autofetch": false
}

If you like CodeLens for tests, keep it on—but it costs CPU on big projects.

4) Stop VS Code from doing “smart” stuff in huge files

#

{
  "editor.largeFileOptimizations": true,
  "files.maxMemoryForLargeFilesMB": 4096
}

5) Extensions: go on a diet

#

In a big repo, too many extensions = death by a thousand cuts.

• Disable anything not essential (Docker/K8s, Python, extra linters, spell-checkers, multiple AI tools, etc.)
• Prefer Disable (Workspace) so other projects aren’t affected.

6) Windows Defender exclusions (often a huge win)

#

Exclude:

• Your repo directory
• Maven cache: C:\Users\<you>\.m2\repository

This can speed up both indexing and mvn runs a lot.

7) Don’t import the whole world into Java

#

If you opened a parent folder containing many repos, VS Code Java will try to be “helpful” and index everything. Open the exact repo root you’re working on.

A good “senior” workflow for big repos in VS Code

#

• Run Spring Boot via terminal/tasks (Maven controls build)
• Use VS Code primarily for navigation + attach debugging
• Keep Java language server memory capped and avoid indexing target/

If you paste:

• your current .vscode/settings.json (or global settings JSON), and
• roughly how many Maven modules, I can suggest a tighter exclude pattern and whether -Xmx2G is enough or you should go 3G.