diff --git a/apps/system/components/ProcessManager/kata-proc-manager/src/sel4bundle/mod.rs b/apps/system/components/ProcessManager/kata-proc-manager/src/sel4bundle/mod.rs
index e3f3085..d5248da 100644
--- a/apps/system/components/ProcessManager/kata-proc-manager/src/sel4bundle/mod.rs
+++ b/apps/system/components/ProcessManager/kata-proc-manager/src/sel4bundle/mod.rs
@@ -292,8 +292,8 @@ impl seL4BundleImpl {
             cap_tcb: CSpaceSlot::new(), // Top-level dup for suspend/resume
             first_page: first_vaddr / PAGE_SIZE,
 
-            affinity: 0, // CPU 0
-            domain: Domain::AppSandbox,
+            affinity: 0,            // CPU 0
+            domain: Domain::System, // TODO(jtgans,sleffler): Figure out how to use this correctly. b/238811077
 
             tcb_name: bundle.app_id.clone(),
             tcb_max_priority: 254, // TODO(sleffler): guess
diff --git a/apps/system/components/kata-os-common/src/scheduling/src/lib.rs b/apps/system/components/kata-os-common/src/scheduling/src/lib.rs
index ece2599..d8a04f2 100644
--- a/apps/system/components/kata-os-common/src/scheduling/src/lib.rs
+++ b/apps/system/components/kata-os-common/src/scheduling/src/lib.rs
@@ -4,11 +4,14 @@
 
 /// Scheduling domains configured for seL4 TCBs.
 ///
-/// Currently we have this setup as a pair of domains, one for the system
-/// components and another for the third party application sandbox.
+/// Currently we have this setup as a single domain for all components, since we
+/// don't want to waste 50% of our time waiting for a mostly idle partition.
+///
+/// TODO: Figure out how to more effectively use these domains of execution, and
+/// how to prevent wasting time in an idle thread for a whole domain when no
+/// TCBs are scheduled there. See also b/238811077.
 ///
 #[derive(Debug, Eq, PartialEq, Copy, Clone)]
 pub enum Domain {
     System = 0,
-    AppSandbox,
 }
diff --git a/easy-settings.cmake b/easy-settings.cmake
index 0e160d8..780ac86 100644
--- a/easy-settings.cmake
+++ b/easy-settings.cmake
@@ -14,5 +14,5 @@ set(SIMULATION ON CACHE BOOL "Whether to build simulate script")
 set(RELEASE OFF CACHE BOOL "Performance optimized build")
 set(UseRiscVBBL OFF CACHE BOOL "Whether to use bbl")
 
-set(KernelNumDomains 2 CACHE STRING "How many scheduling domains to build for")
+set(KernelNumDomains 1 CACHE STRING "How many scheduling domains to build for")
 set(KernelDomainSchedule "${CMAKE_CURRENT_LIST_DIR}/kernel/round_robin_domain.c" CACHE INTERNAL "Domain scheduler algorithm")
diff --git a/kernel/round_robin_domain.c b/kernel/round_robin_domain.c
index a0d10a2..fdea130 100644
--- a/kernel/round_robin_domain.c
+++ b/kernel/round_robin_domain.c
@@ -13,10 +13,19 @@
  * other half. Note that even if there's nothing to run in the third-party
  * application domain, the scheduler will schedule an idle thread to ensure that
  * domain gets it's allocated share of time.
+ *
+ * TODO(jtgans,sleffler): Figure out how to better use these domains for
+ * scheduling applications. We don't really want to use a full 50% duty cycle
+ * for third party applications -- this wastes too much time. See also
+ * b/238811077.
+ *
+ * NOTE: Only a single domain is currently enabled, as per the commented section
+ * below. Any time the below schedule is changed, the number of domains
+ * configured in easy-settings.cmake must also be changed.
  */
 const dschedule_t ksDomSchedule[] = {
     {.domain = 0, .length = 1},  // System domain
-    {.domain = 1, .length = 1},  // Third party application domain
+    /*  {.domain = 1, .length = 1},  // Third party application domain */
 };
 
 const word_t ksDomScheduleLength = sizeof(ksDomSchedule) / sizeof(dschedule_t);