[DPE-4362] fix: alive, restart and alive handling (#109)

src/charm.py

@@ -11,11 +11,19 @@
 from charms.loki_k8s.v0.loki_push_api import LogProxyConsumer
 from charms.prometheus_k8s.v0.prometheus_scrape import MetricsEndpointProvider
 from charms.rolling_ops.v0.rollingops import RollingOpsManager
-from ops import ActiveStatus, InstallEvent, pebble
-from ops.charm import SecretChangedEvent, StorageAttachedEvent, StorageDetachingEvent
-from ops.framework import EventBase
+from ops import (
+    ActiveStatus,
+    EventBase,
+    InstallEvent,
+    SecretChangedEvent,
+    StartEvent,
+    StatusBase,
+    StorageAttachedEvent,
+    StorageDetachingEvent,
+    UpdateStatusEvent,
+    pebble,
+)
 from ops.main import main
-from ops.model import StatusBase
 from ops.pebble import Layer
 
 from core.cluster import ClusterState
@@ -186,9 +194,9 @@ def _on_kafka_pebble_ready(self, event: EventBase) -> None:
         logger.info("Kafka service started")
 
         # service_start might fail silently, confirm with ZK if kafka is actually connected
-        self._on_update_status(event)
+        self.on.update_status.emit()
 
-    def _on_start(self, event: EventBase) -> None:
+    def _on_start(self, event: StartEvent) -> None:
         """Wrapper for start event."""
         self._on_kafka_pebble_ready(event)
 
@@ -257,7 +265,7 @@ def _on_config_changed(self, event: EventBase) -> None:
         if self.model.relations.get(REL_NAME, None) and self.unit.is_leader():
             self.update_client_data()
 
-    def _on_update_status(self, _: EventBase) -> None:
+    def _on_update_status(self, _: UpdateStatusEvent) -> None:
         """Handler for `update-status` events."""
         if not self.upgrade.idle or not self.healthy:
             return

src/core/models.py

@@ -10,10 +10,7 @@
 from charms.zookeeper.v0.client import QuorumLeaderNotFoundError, ZooKeeperManager
 from kazoo.client import AuthFailedError, NoNodeError
 from ops.model import Application, Relation, Unit
-from tenacity import retry
-from tenacity.retry import retry_if_not_result
-from tenacity.stop import stop_after_attempt
-from tenacity.wait import wait_fixed
+from tenacity import retry, retry_if_result, stop_after_attempt, wait_fixed
 from typing_extensions import override
 
 from literals import INTERNAL_USERS, SECRETS_APP, Substrates
@@ -358,12 +355,12 @@ def zookeeper_version(self) -> str:
 
         return zk.get_version()
 
+    # retry to give ZK time to update its broker zNodes before failing
     @retry(
-        # retry to give ZK time to update its broker zNodes before failing
-        wait=wait_fixed(6),
+        wait=wait_fixed(5),
         stop=stop_after_attempt(10),
-        retry_error_callback=(lambda state: state.outcome.result()),  # type: ignore
-        retry=retry_if_not_result(lambda result: True if result else False),
+        retry=retry_if_result(lambda result: result is False),
+        retry_error_callback=lambda _: False,
     )
     def broker_active(self) -> bool:
         """Checks if broker id is recognised as active by ZooKeeper."""