Optimise garbage collection

Note that running garbage collection within a function will not collect
the references this function has, so the queue will actually not be
cleared after being flushed as it was expected. Instead, we use the
option `async` for the GC. This will make this process end its
reductions, and once it has been preempted, it will be scheduled for
garbage collection, and thereafter, a message will be delivered
notifying him of so.

src/mongoose_batch_worker.erl

@@ -53,6 +53,8 @@ handle_cast(Msg, State) ->
 -spec handle_info(flush | term(), state()) -> {noreply, state()}.
 handle_info(flush, State) ->
     {noreply, run_flush(State)};
+handle_info({garbage_collect, asynchronous_gc_triggered, true}, State) ->
+    {noreply, State};
 handle_info(Msg, State) ->
     ?UNEXPECTED_INFO(Msg),
     {noreply, State}.
@@ -100,7 +102,7 @@ run_flush(State = #state{flush_queue = Queue, flush_interval_tref = TRef}) ->
     ?LOG_DEBUG(log_fields(State, #{what => worker_flush})),
     NewState = do_run_flush(State#state{flush_queue = Queue,
                                         flush_interval_tref = undefined}),
-    erlang:garbage_collect(),
+    erlang:garbage_collect(self(), [{async, asynchronous_gc_triggered}, {type, major}]),
     NewState.
 
 cancel_and_flush_timer(undefined) ->