Add config option for customizing scraped metrics

config.yml

@@ -4,10 +4,13 @@ ip_address: 192.168.0.1
 # Connect Box web interface password
 password: WhatEverYourPasswordIs
 
-# The following parameters are optional:
+# All following parameters are optional.
 #exporter:
   # port on which this exporter exposes metrics (default: 9705)
   #port: 9705
 
   # timeout duration for connections to the Connect Box (default: 9)
-  #timeout_seconds: 9
+  #timeout_seconds: 9
+
+  # Customize the family of metrics to scrape. By default, all metrics are scraped.
+  #metrics: [device_status, downstream, upstream, lan_users, temperature]

connectbox_exporter/config.py

@@ -1,17 +1,32 @@
 from pathlib import Path
-import deepmerge
 from typing import Union, Dict
 
+from deepmerge import Merger
 from ruamel.yaml import YAML
 
+from connectbox_exporter.xml2metric import (
+    DEVICE_STATUS,
+    DOWNSTREAM,
+    UPSTREAM,
+    LAN_USERS,
+    TEMPERATURE,
+)
+
 IP_ADDRESS = "ip_address"
 PASSWORD = "password"
 EXPORTER = "exporter"
 PORT = "port"
 TIMEOUT_SECONDS = "timeout_seconds"
+EXTRACTORS = "metrics"
 
 # pick default timeout one second less than the default prometheus timeout of 10s
-DEFAULT_CONFIG = {EXPORTER: {PORT: 9705, TIMEOUT_SECONDS: 9}}
+DEFAULT_CONFIG = {
+    EXPORTER: {
+        PORT: 9705,
+        TIMEOUT_SECONDS: 9,
+        EXTRACTORS: {DEVICE_STATUS, DOWNSTREAM, UPSTREAM, LAN_USERS, TEMPERATURE},
+    }
+}
 
 
 def load_config(config_file: Union[str, Path]) -> Dict:
@@ -24,8 +39,9 @@ def load_config(config_file: Union[str, Path]) -> Dict:
     with open(config_file) as f:
         config = yaml.load(f)
 
-    # merge with defaults
-    config = deepmerge.always_merger.merge(DEFAULT_CONFIG, config)
+    # merge with default config: use 'override' for lists to let users replace extractor setting entirely
+    merger = Merger([(list, "override"), (dict, "merge")], ["override"], ["override"])
+    config = merger.merge(DEFAULT_CONFIG, config)
 
     for param in [IP_ADDRESS, PASSWORD]:
         if not param in config:
@@ -37,4 +53,8 @@ def load_config(config_file: Union[str, Path]) -> Dict:
     if config[EXPORTER][PORT] < 0 or config[EXPORTER][PORT] > 65535:
         raise ValueError(f"Invalid exporter port.")
 
+    if not config[EXPORTER][EXTRACTORS]:
+        raise ValueError("At least one family of metrics needs to be specified.")
+    config[EXPORTER][EXTRACTORS] = sorted(set(config[EXPORTER][EXTRACTORS]))
+
     return config

connectbox_exporter/connectbox_exporter.py

@@ -18,32 +18,27 @@
     EXPORTER,
     PORT,
     TIMEOUT_SECONDS,
+    EXTRACTORS,
 )
 from connectbox_exporter.logger import get_logger, VerboseLogger
-from connectbox_exporter.xml2metric import (
-    TemperatureExtractor,
-    LanUserExtractor,
-    UpstreamStatusExtractor,
-    DownstreamStatusExtractor,
-    DeviceStatusExtractor,
-)
+from connectbox_exporter.xml2metric import get_metrics_extractor
 
 
 class ConnectBoxCollector(object):
     def __init__(
-        self, logger: VerboseLogger, ip_address: str, password: str, timeout: float
+        self,
+        logger: VerboseLogger,
+        ip_address: str,
+        password: str,
+        exporter_config: Dict,
     ):
         self.logger = logger
         self.ip_address = ip_address
         self.password = password
-        self.timeout = timeout
-        self.metric_extractors = [
-            DeviceStatusExtractor(),
-            TemperatureExtractor(),
-            LanUserExtractor(),
-            DownstreamStatusExtractor(),
-            UpstreamStatusExtractor(),
-        ]
+        self.timeout = exporter_config[TIMEOUT_SECONDS]
+
+        extractors = exporter_config[EXTRACTORS]
+        self.metric_extractors = [get_metrics_extractor(e) for e in extractors]
 
     def collect(self):
         # Collect scrape duration and scrape success for each extractor. Scrape success is initialized with False for
@@ -87,6 +82,7 @@ def collect(self):
                     scrape_success[extractor.name] = True
                 except (XMLSyntaxError, AttributeError) as e:
                     # in case of a less serious error, log and continue scraping the next extractor
+                    # TODO make this more useful: log in which extractor the error happened and print the xml at fault
                     self.logger.error(repr(e))
                 except (ConnectionError, Timeout) as e:
                     # in case of serious connection issues, abort and do not try the next extractor
@@ -152,7 +148,7 @@ def main(config_file, verbose):
             logger,
             ip_address=config[IP_ADDRESS],
             password=config[PASSWORD],
-            timeout=exporter_config[TIMEOUT_SECONDS],
+            exporter_config=config[EXPORTER],
         )
     )
 

connectbox_exporter/xml2metric.py

@@ -14,6 +14,12 @@
     StateSetMetricFamily,
 )
 
+DEVICE_STATUS = "device_status"
+DOWNSTREAM = "downstream"
+UPSTREAM = "upstream"
+LAN_USERS = "lan_users"
+TEMPERATURE = "temperature"
+
 
 class XmlMetricsExtractor:
 
@@ -63,7 +69,7 @@ def extract(self, raw_xmls: Dict[int, bytes]) -> Iterable[Metric]:
 class DownstreamStatusExtractor(XmlMetricsExtractor):
     def __init__(self):
         super(DownstreamStatusExtractor, self).__init__(
-            "downstream", {GET.DOWNSTREAM_TABLE, GET.SIGNAL_TABLE}
+            DOWNSTREAM, {GET.DOWNSTREAM_TABLE, GET.SIGNAL_TABLE}
         )
 
     def extract(self, raw_xmls: Dict[int, bytes]) -> Iterable[Metric]:
@@ -152,7 +158,7 @@ def extract(self, raw_xmls: Dict[int, bytes]) -> Iterable[Metric]:
 
 class UpstreamStatusExtractor(XmlMetricsExtractor):
     def __init__(self):
-        super(UpstreamStatusExtractor, self).__init__("upstream", {GET.UPSTREAM_TABLE})
+        super(UpstreamStatusExtractor, self).__init__(UPSTREAM, {GET.UPSTREAM_TABLE})
 
     def extract(self, raw_xmls: Dict[int, bytes]) -> Iterable[Metric]:
         assert len(raw_xmls) == 1
@@ -210,7 +216,7 @@ def extract(self, raw_xmls: Dict[int, bytes]) -> Iterable[Metric]:
 
 class LanUserExtractor(XmlMetricsExtractor):
     def __init__(self):
-        super(LanUserExtractor, self).__init__("lan_users", {GET.LANUSERTABLE})
+        super(LanUserExtractor, self).__init__(LAN_USERS, {GET.LANUSERTABLE})
 
     def extract(self, raw_xmls: Dict[int, bytes]) -> Iterable[Metric]:
         assert len(raw_xmls) == 1
@@ -266,7 +272,7 @@ def extract_client(client, target_metric: GaugeMetricFamily):
 
 class TemperatureExtractor(XmlMetricsExtractor):
     def __init__(self):
-        super(TemperatureExtractor, self).__init__("temperature", {GET.CMSTATE})
+        super(TemperatureExtractor, self).__init__(TEMPERATURE, {GET.CMSTATE})
 
     def extract(self, raw_xmls: Dict[int, bytes]) -> Iterable[Metric]:
         assert len(raw_xmls) == 1
@@ -302,7 +308,7 @@ class ProvisioningStatus(Enum):
 class DeviceStatusExtractor(XmlMetricsExtractor):
     def __init__(self):
         super(DeviceStatusExtractor, self).__init__(
-            "device_status", {GET.GLOBALSETTINGS, GET.CM_SYSTEM_INFO, GET.CMSTATUS}
+            DEVICE_STATUS, {GET.GLOBALSETTINGS, GET.CM_SYSTEM_INFO, GET.CMSTATUS}
         )
 
     def extract(self, raw_xmls: Dict[int, bytes]) -> Iterable[Metric]:
@@ -380,3 +386,24 @@ def extract(self, raw_xmls: Dict[int, bytes]) -> Iterable[Metric]:
             unit="seconds",
             value=uptime_seconds,
         )
+
+
+def get_metrics_extractor(ident: str):
+    """
+    Factory method for metrics extractors.
+    :param ident: metric extractor identifier
+    :return: extractor instance
+    """
+    extractors = {
+        DEVICE_STATUS: DeviceStatusExtractor,
+        DOWNSTREAM: DownstreamStatusExtractor,
+        UPSTREAM: UpstreamStatusExtractor,
+        LAN_USERS: LanUserExtractor,
+        TEMPERATURE: TemperatureExtractor,
+    }
+    if not ident in extractors.keys():
+        raise ValueError(
+            f"Unknown extractor '{ident}', supported are: {','.join(extractors.keys())}"
+        )
+    cls = extractors[ident]
+    return cls()