Use exceptions instead of negative return value

wgkex/worker/mqtt.py

@@ -9,6 +9,7 @@
 from typing import Any, Optional
 
 import paho.mqtt.client as mqtt
+import pyroute2.netlink.exceptions
 
 from wgkex.common import logger
 from wgkex.common.mqtt import (
@@ -231,9 +232,12 @@ def publish_metrics(client: mqtt.Client, topic: str, domain: str) -> None:
             f"Could not get interface name for domain {domain}. Skipping metrics publication"
         )
         return
-    peer_count = get_connected_peers_count(iface)
-    if peer_count < 0:
-        # get_connected_peers_count() encountered some error, don't update metrics
+
+    try:
+        peer_count = get_connected_peers_count(iface)
+    except pyroute2.netlink.exceptions.NetlinkDumpInterrupted:
+        # Handle gracefully, don't update metrics
+        logger.info("Caught NetlinkDumpInterrupted exception while collecting metrics for domain {domain}")
         return
 
     # Publish metrics, retain it at MQTT broker so restarted wgkex broker has metrics right away

wgkex/worker/netlink.py

@@ -218,7 +218,9 @@ def get_connected_peers_count(wg_interface: str) -> int:
 
     Returns:
         The number of peers which have recently seen a handshake.
-        -1 if an internal error has been encountered.
+
+    Raises:
+        NetlinkDumpInterrupted if the interface data has changed while it was being returned by netlink
     """
     three_mins_ago_in_secs = int((datetime.now() - timedelta(minutes=3)).timestamp())
     logger.info("Counting connected wireguard peers for interface %s.", wg_interface)
@@ -227,12 +229,9 @@ def get_connected_peers_count(wg_interface: str) -> int:
             msgs = wg.info(wg_interface)
         except pyroute2.netlink.exceptions.NetlinkDumpInterrupted:
             # Normal behaviour, data has changed while it was being returned by netlink.
-            # Retry once, otherwise return -1 and let the caller handle it.
+            # Retry once, don't catch the exception this time, and let the caller handle it.
             # See https://github.com/svinota/pyroute2/issues/874
-            try:
-                msgs = wg.info(wg_interface)
-            except pyroute2.netlink.exceptions.NetlinkDumpInterrupted:
-                return -1
+            msgs = wg.info(wg_interface)
 
         logger.debug("Got infos for connected peers: %s.", msgs)
         count = 0