perf: reduce tokio worker thread blocking to improve responsiveness on low-core machines

- Switch memdb durability from Full to Batch(100) with 10s flush worker
  - Offload BlobManager fjall writes to spawn_blocking
  - Wrap Tantivy commit operations in block_in_place
  - Flush memdb WAL on graceful shutdown

crates/server/src/lib.rs

@@ -28,6 +28,7 @@ use bichon_core::{
     cache::imap::task::SYNC_TASKS,
     common::{rustls::BichonTls, signal::SignalManager},
     context::{executors::BichonContext, Initialize},
+    database::manager::DB_MANAGER,
     error::{code::ErrorCode, BichonResult},
     logger,
     migrate::check_data_status,
@@ -116,6 +117,7 @@ pub async fn run() -> BichonResult<()> {
     ENVELOPE_MANAGER.shutdown().await;
     ATTACHMENT_MANAGER.shutdown().await;
     BLOB_MANAGER.shutdown().await;
+    DB_MANAGER.flush();
     info!("Bichon server stopped.");
     Ok(())
 }