Use pooled Netty allocator in tests by default (elastic#100877)

We mostly run our tests with less than 1G of heap per JVM. This means that
we will use the unpooled Netty allocator in most tests, losing us a lot of
leak coverage in internal cluster tests (mostly for inbound buffers).
Unless otherwise specified by tests, we should force the use of our standard
allocator by default to get a higher chance of catching leaks in internalClusterTests
in particular.

test/framework/src/main/java/org/elasticsearch/test/ESTestCase.java

@@ -323,6 +323,11 @@ private static void setTestSysProps() {
 
         // Enable Netty leak detection and monitor logger for logged leak errors
         System.setProperty("io.netty.leakDetection.level", "paranoid");
+        if (System.getProperty("es.use_unpooled_allocator") == null) {
+            // unless explicitly forced to unpooled, always test with the pooled allocator to get the best possible coverage from Netty's
+            // leak detection which does not cover simple unpooled heap buffers
+            System.setProperty("es.use_unpooled_allocator", "false");
+        }
 
         // We have to disable setting the number of available processors as tests in the same JVM randomize processors and will step on each
         // other if we allow them to set the number of available processors as it's set-once in Netty.