fix: update the simple profile to avoid Container.can_connect (#1927)

Using `Container.can_connect` as a guard for Pebble is a poor pattern,
because it introduces a race condition. We recommend that charms use the
"easier to ask forgiveness than permission" pattern by executing the
container calls and catching any errors that occur. This PR updates the
`simple` profile to use that pattern.

There are two other small changes:
* The message in the blocked status was intending to use an f-string,
but was just a regular string, so would have always shown the literal
text "{log_level}" rather than the value of that variable.
* `WaitingStatus` should be used when waiting on an integrated charm,
not the local charm. `MaintenanceStatus` is the correct class to use
when waiting on the local charm.

charmcraft/templates/init-simple/src/charm.py.j2

@@ -64,18 +64,18 @@ class {{ class_name }}(ops.CharmBase):
         if log_level in VALID_LOG_LEVELS:
             # The config is good, so update the configuration of the workload
             container = self.unit.get_container("httpbin")
-            # Verify that we can connect to the Pebble API in the workload container
-            if container.can_connect():
-                # Push an updated layer with the new config
+            # Push an updated layer with the new config
+            try:
                 container.add_layer("httpbin", self._pebble_layer, combine=True)
                 container.replan()
-
-                logger.debug("Log level for gunicorn changed to '%s'", log_level)
-                self.unit.status = ops.ActiveStatus()
-            else:
+            except ops.pebble.ConnectionError:
                 # We were unable to connect to the Pebble API, so we defer this event
-                event.defer()
                 self.unit.status = ops.MaintenanceStatus("waiting for Pebble API")
+                event.defer()
+                return
+
+            logger.debug("Log level for gunicorn changed to '%s'", log_level)
+            self.unit.status = ops.ActiveStatus()
         else:
             # In this case, the config option is bad, so block the charm and notify the operator.
             self.unit.status = ops.BlockedStatus(f"invalid log level: '{log_level}'")