Obsidian MCP Server

# Cache Warmup Pre-load caches on startup for immediate performance. ## Purpose Cache Warmup pre-populates caches and builds indexes on application startup, ensuring optimal performance from the first request rather than warming up gradually. ## How It Works ```mermaid flowchart TD A[Server Start] --> B{Warmup enabled?} B -->|No| C[Normal startup] B -->|Yes| D[List vault files] D --> E{File count > limit?} E -->|Yes| F[Limit to warmupFileCount] E -->|No| G[Process all files] F --> G G --> H[Read file content] H --> I[Parse frontmatter] I --> J[Update content cache] J --> K[Update inverted index] K --> L[Update path trie] L --> M{More files?} M -->|Yes| H M -->|No| N[Mark as warmed up] N --> O[Ready for requests] ``` ## Configuration | Variable | Type | Default | Description | | ------------------------------ | --------- | ------- | ------------------------------- | | `OBSIDIAN_ENABLE_CACHE_WARMUP` | `boolean` | `false` | Enable warmup on startup | | `OBSIDIAN_WARMUP_FILE_COUNT` | `number` | `1000` | Maximum files to pre-load | ## Warmup Sequence 1. **List Files**: Scan vault for all markdown files 2. **Prioritize**: Sort by modification time (most recent first) 3. **Limit**: Take only `warmupFileCount` files 4. **Read**: Load file content into content cache 5. **Parse**: Extract frontmatter and inline tags 6. **Index**: Add to inverted index (if enabled) 7. **Trie**: Add to path trie (if enabled) 8. **Complete**: Set `isWarmedUp` flag ## Startup Time Impact | Vault Size | Without Warmup | With Warmup (1000 files) | | ------------- | -------------- | ------------------------ | | 100 files | <100ms | ~500ms | | 1,000 files | <100ms | ~2-3s | | 10,000 files | <100ms | ~5-10s (limited) | ## First Request Performance | Scenario | Without Warmup | With Warmup | | ------------------ | -------------- | ----------- | | First search | Cold (slow) | Warm (fast) | | First file read | Disk I/O | Memory | | First glob match | Trie building | Instant | ## When to Use Enable warmup when: - Server restarts are infrequent - First-request latency is critical - Vault size is manageable (< 5000 files) Skip warmup when: - Fast startup is critical - Running in serverless environment - Vault is very large (> 10000 files) ## Warmup Status Check if warmup completed via [get_cache_stats](../tools/get_cache_stats.md): ```json { "isWarmedUp": true } ``` ## Persistence Integration When persistent cache is also enabled: ```mermaid flowchart TD A[Server Start] --> B{Persistent cache exists?} B -->|Yes| C[Load from LMDB] B -->|No| D[Run warmup] C --> E{Still valid?} E -->|Yes| F[Skip warmup] E -->|No| D D --> G[Save to LMDB] F --> H[Ready] G --> H ``` This provides the best of both worlds: - Fast restarts from persistent cache - Full warmup only on first run ## Source Files - [`src/services/fileSystem.ts`](../../src/services/fileSystem.ts) - `warmCache()` method ## Related Features - [Persistent Cache](persistent-cache.md) - Stores warmed data - [Content Cache](content-cache.md) - Populated during warmup - [Inverted Index](inverted-index.md) - Built during warmup - [Path Trie](path-trie.md) - Built during warmup