[mem] Flatten memory checkers into composable per-pool invariant checks

python/sglang/srt/disaggregation/decode.py

@@ -1191,7 +1191,7 @@ def event_loop_normal_disagg_decode(self: Scheduler):
                 self.process_batch_result(batch, result)
             else:
                 # When the server is idle, do self-check and re-init some states
-                self.self_check_during_idle()
+                self.on_idle()
 
             # Update last_batch
             self.last_batch = batch
@@ -1224,7 +1224,7 @@ def event_loop_overlap_disagg_decode(self: Scheduler):
                 tmp_batch, tmp_result = self.result_queue.popleft()
                 self.process_batch_result(tmp_batch, tmp_result)
             elif batch is None:
-                self.self_check_during_idle()
+                self.on_idle()
 
             # Run sample of the current batch
             # It depends on the result of the last batch (e.g., grammar), so we run it after the last batch is processed.

python/sglang/srt/disaggregation/prefill.py

@@ -409,7 +409,7 @@ def event_loop_normal_disagg_prefill(self: Scheduler) -> None:
                 result = self.run_batch(batch)
                 self.process_batch_result(batch, result)
             else:
-                self.self_check_during_idle()
+                self.on_idle()
 
             self.process_disagg_prefill_inflight_queue()
 
@@ -448,7 +448,7 @@ def event_loop_overlap_disagg_prefill(self: Scheduler) -> None:
                 self.process_batch_result(tmp_batch, tmp_result)
             elif batch is None:
                 # When the server is idle, do self-check and re-init some states
-                self.self_check_during_idle()
+                self.on_idle()
 
             self.process_disagg_prefill_inflight_queue()
 

python/sglang/srt/managers/scheduler.py

@@ -1370,7 +1370,7 @@ def event_loop_normal(self):
                 self.process_batch_result(batch, result)
             else:
                 # When the server is idle, do self-check and re-init some states.
-                self.self_check_during_idle()
+                self.on_idle()
 
             # Update last_batch
             self.last_batch = batch
@@ -1420,7 +1420,7 @@ def pop_and_process():
                     pop_and_process()
             elif batch is None:
                 # When the server is idle, do self-check and re-init some states
-                self.self_check_during_idle()
+                self.on_idle()
 
             # Run sample of the current batch
             # It depends on the result of the last batch (e.g., grammar), so we run it after the last batch is processed.

python/sglang/srt/managers/scheduler_pp_mixin.py

@@ -142,7 +142,7 @@ def event_loop_pp(self: Scheduler):
 
             # When the server is idle, self-check and re-init some states
             if server_is_idle:
-                self.self_check_during_idle()
+                self.on_idle()
 
     @DynamicGradMode()
     def event_loop_pp_disagg_prefill(self: Scheduler):
@@ -318,7 +318,7 @@ def event_loop_pp_disagg_prefill(self: Scheduler):
 
             # When the server is idle, self-check and re-init some states
             if server_is_idle and len(self.disagg_prefill_inflight_queue) == 0:
-                self.self_check_during_idle()
+                self.on_idle()
 
     @DynamicGradMode()
     def event_loop_pp_disagg_decode(self: Scheduler):
@@ -508,7 +508,7 @@ def event_loop_pp_disagg_decode(self: Scheduler):
                 queue_size += len(self.decode_offload_manager.ongoing_offload)
 
             if server_is_idle and queue_size == 0:
-                self.self_check_during_idle()
+                self.on_idle()
 
     def init_pp_loop_state(self: Scheduler):
         self.pp_loop_size: int = self.pp_size + self.server_args.pp_async_batch_depth