use default snapshot_threshold of 10000

Signed-off-by: PeaBrane <yanrpei@gmail.com>

Author: PeaBrane

components/frontend/src/dynamo/frontend/main.py

@@ -152,8 +152,8 @@ def parse_args():
     parser.add_argument(
         "--router-snapshot-threshold",
         type=int,
-        default=None,
-        help="KV Router: Number of messages in stream before triggering a snapshot. If not set, snapshots are disabled.",
+        default=10000,
+        help="KV Router: Number of messages in stream before triggering a snapshot. Defaults to 10000.",
     )
     parser.add_argument(
         "--router-persist-states",

docs/architecture/kv_cache_routing.md

@@ -27,7 +27,7 @@ The main KV-aware routing arguments:
 
 - `--router-reset-states`/`--router-persist-states`: Controls whether the router state is reset on startup. When `--router-reset-states` is used (default), the router clears both the JetStream event stream and NATs object store, starting with a fresh state. When `--router-persist-states` is used, the router retains existing state from previous runs, downloading any available snapshot from NATs object store and continuing to consume events from where it left off. This enables routers to maintain KV cache awareness across restarts. **Note**: State persistence is only available when `--use-kv-events` is enabled (default). When using `--no-kv-events` with `ApproxKvIndexer`, state persistence is not supported.
 
-- `--router-snapshot-threshold`: Sets the number of messages in the JetStream before triggering a snapshot. When the message count exceeds this threshold, a router will attempt to purge acknowledged messages from the stream and create a snapshot of the current radix tree state in NATs object store. If not set, snapshots are disabled. This helps manage stream size and provides faster initialization for routers that restart.
+- `--router-snapshot-threshold`: Sets the number of messages in the JetStream before triggering a snapshot. When the message count exceeds this threshold, a router will attempt to purge acknowledged messages from the stream and create a snapshot of the current radix tree state in NATs object store. Defaults to 10000. This helps manage stream size and provides faster initialization for routers that restart.
 
 ## Architecture
 

lib/bindings/python/rust/llm/entrypoint.rs

@@ -42,7 +42,7 @@ impl KvRouterConfig {
 #[pymethods]
 impl KvRouterConfig {
     #[new]
-    #[pyo3(signature = (overlap_score_weight=1.0, router_temperature=0.0, use_kv_events=true, router_replica_sync=false, router_snapshot_threshold=None, router_reset_states=true))]
+    #[pyo3(signature = (overlap_score_weight=1.0, router_temperature=0.0, use_kv_events=true, router_replica_sync=false, router_snapshot_threshold=10000, router_reset_states=true))]
     fn new(
         overlap_score_weight: f64,
         router_temperature: f64,

lib/llm/src/kv_router.rs

@@ -121,7 +121,7 @@ impl Default for KvRouterConfig {
             use_kv_events: true,
             router_replica_sync: false,
             max_num_batched_tokens: 8192,
-            router_snapshot_threshold: None,
+            router_snapshot_threshold: Some(10000),
             router_reset_states: true,
         }
     }