logging: log to console

As RabbitMQ is running under Pebble inside of a container, make
sure all log output goes to the console so it can easily be accessed
and collecting into observability solutions.

This requires that we bypass the helper wrapper script which will
always redirect STDOUT and STDERR to a log file.

(cherry picked from commit acf91f9)

src/charm.py

@@ -218,14 +218,18 @@ def _check_rmq_running():
 
     def _rabbitmq_layer(self) -> dict:
         """Pebble layer definition for RabbitMQ."""
+        # NOTE(jamespage)
+        # Use the full path to the rabbitmq-server binary
+        # rather than the helper wrapper script to avoid
+        # redirection of console output to a log file.
         return {
             "summary": "RabbitMQ layer",
             "description": "pebble config layer for RabbitMQ",
             "services": {
                 RABBITMQ_SERVICE: {
                     "override": "replace",
                     "summary": "RabbitMQ Server",
-                    "command": "rabbitmq-server",
+                    "command": "/usr/lib/rabbitmq/bin/rabbitmq-server",
                     "startup": "enabled",
                     "user": RABBITMQ_USER,
                     "group": RABBITMQ_GROUP,
@@ -588,6 +592,10 @@ def _render_and_push_rabbitmq_conf(self) -> None:
 cluster_partition_handling = autoheal
 
 queue_master_locator = min-masters
+
+# Log to console for pod logging
+log.console = true
+log.file = false
 """
         logger.info("Pushing new rabbitmq.conf")
         container.push(

tests/unit/test_charm.py

@@ -73,7 +73,7 @@ def test_rabbitmq_pebble_ready(self):
                 "rabbitmq": {
                     "override": "replace",
                     "summary": "RabbitMQ Server",
-                    "command": "rabbitmq-server",
+                    "command": "/usr/lib/rabbitmq/bin/rabbitmq-server",
                     "startup": "enabled",
                     "user": "rabbitmq",
                     "group": "rabbitmq",