diff --git a/repos/os/run/monitor.run b/repos/os/run/monitor.run
index c53186fe35..1f5911d7f2 100644
--- a/repos/os/run/monitor.run
+++ b/repos/os/run/monitor.run
@@ -42,7 +42,7 @@ install_config {
-
+
@@ -60,7 +60,7 @@ install_config {
-
+
diff --git a/repos/os/src/test/monitor/main.cc b/repos/os/src/test/monitor/main.cc
index ad4b550f31..c249d9ee96 100644
--- a/repos/os/src/test/monitor/main.cc
+++ b/repos/os/src/test/monitor/main.cc
@@ -124,17 +124,16 @@ struct Test::Main
/*
* Dimensioning of the buffer for one round trip:
*
- * The terminal_crosslink component uses a buffer of 4 KiB and
- * the debug monitor limits the 'm' command response to 2 KiB to leave
- * enough space for asynchronous notifications and protocol overhead.
- * GDB's 'm' command encodes memory as hex, two characters per byte.
- * Hence, a dump of max. 1 KiB is currently possible.
- *
* The most effective way to optimize the throughput would be to
- * increase the terminal-crosslink's buffer size, reducing the number
- * of round trips.
+ * increase the terminal-crosslink's buffer size and the maximum GDB
+ * response size of the debug monitor, reducing the number of round
+ * trips.
+ *
+ * GDB's 'm' command encodes memory as hex, two characters per byte.
+ * Hence, a dump of max. 1/2 of the configured GDB response size is
+ * possible.
*/
- char buffer[1024] { };
+ char buffer[4096] { };
uint64_t const start_us = timer.elapsed_us();
uint64_t now_us = start_us;