Approve failsafe with temp file

config.sample.toml

@@ -3,8 +3,13 @@ source_collector_dir = "path/to/source_collector_dir/"
 host_modifier_dir = "path/to/host_modifier_dir/"
 db_uri = "dbname='zac' user='zabbix' host='localhost' password='secret' port=5432 connect_timeout=2"
 log_level = "DEBUG"
+# Health status for each ZAC process.
 health_file = "/tmp/zac_health.json"
+# File containing hostnames of hosts to add/remove when failsafe is reached.
 failsafe_file = "/tmp/zac_failsafe.json"
+# File to signal manual approval of adding/removing hosts when failsafe is reached.
+# The file is automatically removed after changes are made.
+failsafe_ok_file = "/tmp/zac_failsafe_ok"
 
 [zabbix]
 map_dir = "path/to/map_dir/"

zabbix_auto_config/models.py

@@ -82,11 +82,12 @@ class ZacSettings(ConfigBaseModel):
     source_collector_dir: str
     host_modifier_dir: str
     db_uri: str
-    health_file: Optional[Path] = None
     log_level: int = Field(logging.DEBUG, description="The log level to use.")
+    health_file: Optional[Path] = None
     failsafe_file: Optional[Path] = None
+    failsafe_ok_file: Optional[Path] = None
 
-    @field_validator("failsafe_file", "health_file", mode="after")
+    @field_validator("health_file", "failsafe_file", "failsafe_ok_file", mode="after")
     @classmethod
     def _validate_file_path(cls, v: Optional[Path], info: ValidationInfo) -> Optional[Path]:
         if v is None:
@@ -283,4 +284,4 @@ class HostActions(BaseModel):
 
     def write_json(self, path: Path) -> None:
         """Writes a JSON serialized representation of self to a file."""
-        utils.write_file(path, self.model_dump_json(indent=2))
+        utils.write_file(path, self.model_dump_json(indent=2))

zabbix_auto_config/processing.py

@@ -733,6 +733,22 @@ def set_tags(self, zabbix_host, tags):
         else:
             logging.info("DRYRUN: Setting tags (%s) on host: '%s' (%s)", tags, zabbix_host["host"], zabbix_host["hostid"])
 
+    def handle_failsafe_limit(self, to_add: List[str], to_remove: List[str]) -> None:
+        """Handles situations where the number of hosts to add/remove exceeds the failsafe.
+
+        If a failsafe OK file exists, the method will remove it and proceed with the changes.
+        Otherwise, it will write the list of hosts to add and remove to a failsafe file and
+        raise a ZACException."""
+        if self._check_failsafe_ok_file():
+            return
+        self.write_failsafe_hosts(to_add, to_remove)
+        logging.warning(
+            "Too many hosts to change (failsafe=%d). Remove: %d, Add: %d. Aborting",
+            self.config.failsafe,
+            len(to_remove),
+            len(to_add),
+        )
+        raise exceptions.ZACException("Failsafe triggered")
 
     def write_failsafe_hosts(self, to_add: List[str], to_remove: List[str]) -> None:
         if not self.settings.zac.failsafe_file:
@@ -747,6 +763,25 @@ def write_failsafe_hosts(self, to_add: List[str], to_remove: List[str]) -> None:
             self.settings.zac.failsafe_file,
         )
 
+    def _check_failsafe_ok_file(self) -> bool:
+        if not self.settings.zac.failsafe_ok_file:
+            return False
+        if not self.settings.zac.failsafe_ok_file.exists():
+            logging.info(
+                "Failsafe OK file %s does not exist. Create it to approve changes.",
+                self.settings.zac.failsafe_ok_file,
+            )
+            return False
+        try:
+            self.settings.zac.failsafe_ok_file.unlink()
+        except OSError as e:
+            logging.error(
+                "Failsafe cannot be approved. Unable to delete failsafe OK file: %s", e
+            )
+            return False
+        logging.info("Failsafe OK file exists. Proceeding with changes.")
+        return True
+
     def do_update(self):
         with self.db_connection, self.db_connection.cursor() as db_cursor:
             db_cursor.execute(f"SELECT data FROM {self.db_hosts_table} WHERE data->>'enabled' = 'true'")
@@ -795,10 +830,11 @@ def do_update(self):
         logging.debug("Only in db: %s", " ".join(hostnames_to_add[:10]))
         logging.debug("In both: %d", len(hostnames_in_both))
 
-        if len(hostnames_to_remove) > self.config.failsafe or len(hostnames_to_add) > self.config.failsafe:
-            logging.warning("Too many hosts to change (failsafe=%d). Remove: %d, Add: %d. Aborting", self.config.failsafe, len(hostnames_to_remove), len(hostnames_to_add))
-            self.write_failsafe_hosts(hostnames_to_add, hostnames_to_remove)
-            raise exceptions.ZACException("Failsafe triggered")
+        if (
+            len(hostnames_to_remove) > self.config.failsafe
+            or len(hostnames_to_add) > self.config.failsafe
+        ):
+            self.handle_failsafe_limit(hostnames_to_add, hostnames_to_remove)
 
         for hostname in hostnames_to_remove:
             if self.stop_event.is_set():