Hibernate 2 metrics gc processes

It was observed that `rabbit_core_metrics_gc` and
`rabbit_stream_metrics_gc` processes can grow to several MBs of
memory (probably because fetching the list of all queues). As they
execute infrequently (every 2 minutes by default) it can save some
memory to hibernate them in-between (similar to other similar
processes).

Author: gomoripeti

deps/rabbit/src/rabbit_core_metrics_gc.erl

@@ -6,6 +6,8 @@
 %%
 -module(rabbit_core_metrics_gc).
 
+-behaviour(gen_server).
+
 -record(state, {timer,
                 interval
                }).
@@ -17,7 +19,7 @@
 -spec start_link() -> rabbit_types:ok_pid_or_error().
 
 start_link() ->
-    gen_server:start_link({local, ?MODULE}, ?MODULE, [], []).
+    gen_server:start_link({local, ?MODULE}, ?MODULE, [], [{hibernate_after, 0}]).
 
 init(_) ->
     Interval = rabbit_misc:get_env(rabbit, core_metrics_gc_interval, 120000),

deps/rabbitmq_stream/src/rabbit_stream_metrics_gc.erl

@@ -32,7 +32,7 @@
 
 -spec start_link() -> rabbit_types:ok_pid_or_error().
 start_link() ->
-    gen_server:start_link({local, ?MODULE}, ?MODULE, [], []).
+    gen_server:start_link({local, ?MODULE}, ?MODULE, [], [{hibernate_after, 0}]).
 
 init(_) ->
     Interval =