diff --git a/changes/bug32108 b/changes/bug32108
new file mode 100644
index 0000000000..2806fa3e5d
--- /dev/null
+++ b/changes/bug32108
@@ -0,0 +1,8 @@
+  o Major bugfixes (relay):
+    - Relays now respect their AccountingMax bandwidth again. When relays
+      entered "soft" hibernation (which typically starts when we've hit
+      90% of our AccountingMax), we had stopped checking whether we should
+      enter hard hibernation. Soft hibernation refuses new connections and
+      new circuits, but the existing circuits can continue, meaning that
+      relays could have exceeded their configured AccountingMax. Fixes
+      bug 32108; bugfix on 0.4.0.1-alpha.
diff --git a/src/core/mainloop/mainloop.c b/src/core/mainloop/mainloop.c
index 18e87fa87a..193df61d01 100644
--- a/src/core/mainloop/mainloop.c
+++ b/src/core/mainloop/mainloop.c
@@ -1397,7 +1397,7 @@ STATIC periodic_event_item_t periodic_events[] = {
   /* This is a legacy catch-all callback that runs once per second if
    * we are online and active. */
   CALLBACK(second_elapsed, NET_PARTICIPANT,
-           FL(NEED_NET)|FL(RUN_ON_DISABLE)),
+           FL(RUN_ON_DISABLE)),
 
   /* XXXX Do we have a reason to do this on a callback? Does it do any good at
    * all?  For now, if we're dormant, we can let our listeners decay. */