Merge pull request 'Version 2.0' (#4) from dev into main

Reviewed-on: #4
v.2.0:
Refactored all metrics
Added labels
Added self measured metrics
Updated README.md

README.md

@@ -14,6 +14,8 @@ Create configurable lightweight application to collect some metrics
 - hot reload metrics if configuration changed
 - could be used as [regular application](#StartRegular), as [systemctl service](#StartService) or a [Docker application](#StartDocker)
 - supports JSON, PROPERTIES and YAML configuration formats
+- [internal metrics](#Internal) to show how time spent on update every other metrics
+- [metrics labels](#Labels) support
 
 ### 📌 Using
 _To use specific configuration format the `app_config.CONFIG_FILE_NAME` variable need to be changed. The default config file format is **JSON**_
@@ -80,6 +82,12 @@ The Default Application config is (no custom metrics are configured):
 - `name` - parameter used in every metric to identify it. **Required**.
 - `interval` - time interval in seconds the metric will be updated. **Required**.
 
+#### Metrics Labels<a id='Labels' />
+From version 2.0 the Application supports Labels. See [Metrics Names](MetricName) for details.
+
+#### Internal metrics<a id='Internal' />
+From version 2.0 the Application supports internal metrics to collect time. See [Metrics Names](MetricName) for details.
+
 #### Disk (or mount point) Metrics<a id='DiscMetrics' />
 **_Monitors the Mount Point's sizes: `total`, `used`, `free` space in bytes_**
 ```json
@@ -187,11 +195,21 @@ The Default Application config is (no custom metrics are configured):
 - `args` - CLI arguments to be provided to the command
 In example above the metric will return integer value 3.
 
-<a id='MetricName' />**The metric name creates as follows:**
-- uses Metric Prefix, actually `das_`
-- uses `metric_text` given to every metric while it creating
-- uses `instance_prefix` given in metric configuration
-- uses `name` given in metric configuration
+<a id='MetricName' />**The metric names:**
+From version 2.0 there are following metric names used
+- `das_collect_time_ms` - Total time spent collecting metrics  in milliseconds; Labels: **name**, Total time spent collecting metrics [name] on [server] in milliseconds
+- `das_disk_bytes` - Bytes (total, used, free) on (mount_point) for (server); Labels: **total, used, free, mount_point, server**
+- `das_service_health` - Service health; Labels **name, url, method, server**
+- `das_rest_value` - Remote REST API Value; Labels **name, url, method, server**
+- `das_shell_value` - Shell Value; Labels: **name, command, server**
+- `das_host_available` - Host availability; Labels **name, ip, server**
+- `das_net_interface_bytes` - Network Interface bytes; Labels: **name, server, metric**=(sent|receive)
+- `das_exporter` - Exporter Uptime for **server** in seconds
+- `das_uptime_seconds` - System uptime on **server**
+- `das_cpu_percent` - CPU used percent on **server**
+- `das_memory_percent` - Memory used percent on **server**
+- `das_ChassisTemperature_current` - Current Chassis Temperature overall on **server** 
+- `das_CpuTemperature_current` - Current CPU Temperature overall on **server**
 **Note:** there are no doubles in metrics names supported by Prometheus. If so the exception occurs ant the application will be stopped.
 
 ### 🚀 Launching the application

app_config.py

@@ -1,5 +1,6 @@
 import os
 
+APP_VERSION="2.0"
 SCRIPT_PATH = os.path.dirname(__file__)
 CONFIGS_DIR = SCRIPT_PATH + "/configs"
 CONFIG_FILE_NAME = CONFIGS_DIR + "/config.json"

config_file.py

@@ -16,7 +16,7 @@ def read_config(name):
         raise Exception("Wrong file type")
 
 def read_json(name):
-    with open(name, 'r') as f:
+    with open(name, 'r', encoding='utf-8') as f:
         j_conf = json.load(f)
         conf = {}
         for key, value in j_conf.items():
@@ -25,7 +25,7 @@ def read_json(name):
 
 def read_prop(filepath, sep='=', comment_char='#'):
     conf = {}
-    with open(filepath, "rt") as f:
+    with open(filepath, "rt", encoding='utf-8') as f:
         for line in f:
             l = line.strip()
             if l and not l.startswith(comment_char):
@@ -37,7 +37,7 @@ def read_prop(filepath, sep='=', comment_char='#'):
 
 def read_yaml(name):
     conf = {}
-    with open(name, 'r') as f:
+    with open(name, 'r', encoding='utf-8') as f:
         y_conf = yaml.safe_load(f)
         for key, value in y_conf.items():
             conf[key] = value

main.py

@@ -59,6 +59,7 @@ def is_need_to_reload_config():
 
 def print_config_info_debug():
     print('-=: Debug Mode :=-')
+    print(f'\tAPP_VERSION={app_config.APP_VERSION}')
     print(f'\tSCRIPT_PATH={app_config.SCRIPT_PATH}')
     print(f'\tCONFIGS_DIR={app_config.CONFIGS_DIR}')
     print(f'\tCONFIG_FILE_NAME={app_config.CONFIG_FILE_NAME}')
@@ -74,7 +75,7 @@ def print_config_info_debug():
     print(f'\tIS_PRINT_INFO={app_config.IS_PRINT_INFO}')
 
 def main():
-    print("-=: Collector started :=-")
+    print(f'-=: Collector started (version {app_config.APP_VERSION}) :=-')
     if os.path.isfile(app_config.CONFIG_FILE_NAME):
         config = read_app_config()
         parse_config(config)

metrics/DataStructures.py

@@ -11,32 +11,57 @@ ENUM_UP_DN_STATES = ['up', 'dn']
 def get_metric(name):
     return REGISTRY._names_to_collectors.get(name)
 
-def get_gauge_metric(metric_name, descr):
+def get_gauge_metric(metric_name, descr, labels=None):
+    if labels is None:
+        labels = []
     metric = get_metric(metric_name)
     if metric is None:
-        metric = Gauge(metric_name, descr)
+        if labels:
+            metric = Gauge(metric_name, descr, labelnames=labels)
+        else:
+            metric = Gauge(metric_name, descr)
     return metric
 
-def get_counter_metric(metric_name, descr):
+def get_counter_metric(metric_name, descr, labels=None):
     metric = get_metric(metric_name)
     if metric is None:
-        metric = Counter(metric_name, descr)
+        if labels:
+            metric = Counter(metric_name, descr, labelnames=labels)
+        else:
+            metric = Counter(metric_name, descr)
     return metric
 
-def get_enum_metric(metric_name, descr, states):
+def get_enum_metric(metric_name, descr, states, labels=None):
     metric = get_metric(metric_name)
     if metric is None:
-        metric = Enum(metric_name, descr, states=states)
+        if labels:
+            metric = Enum(metric_name, descr, states=states, labelnames=labels)
+        else:
+            metric = Enum(metric_name, descr, states=states)
     return metric
 
+def get_time_millis():
+    return round(time.time() * 1000)
+
 
 class AbstractData:
-    METRIC_NAME_PREFIX = 'das_'
+    g_collect: Gauge
     def __init__(self, name, interval, prefix=''):
         self.name = name
         self.interval = interval
         self.instance_prefix = prefix
         self.updated_at = int(time.time())
+        self.g_collect = get_gauge_metric('das_collect_time_ms',
+                                          'Total time spent collecting metrics [name] on [server] in milliseconds',
+                                          ['server', 'name'])
+        self.g_collect.labels(server=prefix, name=name)
+        self.g_collect.labels(server=prefix, name=name)
+        self.g_collect.labels(server=prefix, name=name)
+        self.g_collect.labels(server=prefix, name=name)
+        self.g_collect.labels(server=prefix, name=name)
+        self.g_collect.labels(server=prefix, name=name)
+        self.g_collect.labels(server=prefix, name=name)
+        self.g_collect.labels(server=prefix, name=name)
 
     def set_update_time(self):
         self.updated_at = int(time.time())
@@ -44,11 +69,8 @@ class AbstractData:
     def is_need_to_update(self):
         return self.updated_at + self.interval <= int(time.time())
 
-    def get_metric_name(self, metric_text, name):
-        return (self.METRIC_NAME_PREFIX +
-                metric_text + '_' +
-                (self.instance_prefix + '_' if self.instance_prefix else '') +
-                name)
+    def set_collect_time(self, value=0):
+        self.g_collect.labels(server=self.instance_prefix, name=self.name).set(value)
 
     def print_trigger_info(self):
         if app_config.IS_PRINT_INFO:
@@ -56,24 +78,27 @@ class AbstractData:
 
 
 class DiskData(AbstractData):
-    g_total: Gauge
-    g_used: Gauge
-    g_free: Gauge
+    g_all: Gauge
     def __init__(self, mount_point='/', total=0, used=0, free=0, interval=60, name='', prefix=''):
         super().__init__(name, interval, prefix)
         self.mount_point = mount_point
         self.total = total
         self.used = used
         self.free = free
-        self.g_total = get_gauge_metric(self.get_metric_name('disk_total_bytes', name), 'Total bytes on disk')
-        self.g_used = get_gauge_metric(self.get_metric_name('disk_used_bytes', name), 'Used bytes on disk')
-        self.g_free = get_gauge_metric(self.get_metric_name('disk_free_bytes', name), 'Free bytes on disk')
+        self.g_all = get_gauge_metric('das_disk_bytes',
+                                      'Bytes [total, used, free] on [mount_point] for [server]',
+                                      ['name', 'mount', 'server', 'metric'])
+        self.g_all.labels(name=name, mount=mount_point, server=self.instance_prefix, metric='total')
+        self.g_all.labels(name=name, mount=mount_point, server=self.instance_prefix, metric='used')
+        self.g_all.labels(name=name, mount=mount_point, server=self.instance_prefix, metric='free')
         self.set_data(total, used, free)
 
     def set_data(self, total, used, free):
-        self.g_total.set(total)
-        self.g_used.set(used)
-        self.g_free.set(free)
+        time_ms = get_time_millis()
+        self.g_all.labels(name=self.name, mount=self.mount_point, server=self.instance_prefix, metric='total').set(total)
+        self.g_all.labels(name=self.name, mount=self.mount_point, server=self.instance_prefix, metric='used').set(used)
+        self.g_all.labels(name=self.name, mount=self.mount_point, server=self.instance_prefix, metric='free').set(free)
+        self.set_collect_time(get_time_millis() - time_ms)
         self.set_update_time()
         self.print_trigger_info()
 
@@ -91,13 +116,17 @@ class HealthData(AbstractData):
         self.user = user
         self.password = password
         self.headers = headers
-        metric_name = self.get_metric_name('service_health', name)
-        self.e_state = get_enum_metric(metric_name, 'Service health', ENUM_UP_DN_STATES)
-        self.set_status(is_up)
+        self.e_state = get_enum_metric('das_service_health',
+                                       'Service [name, url, method, server] health',
+                                       ENUM_UP_DN_STATES,['name', 'url', 'method', 'server'])
+        self.e_state.labels(name=name, url=url, method=method, server=self.instance_prefix)
+        self.set_data(is_up)
 
-    def set_status(self, is_up):
+    def set_data(self, is_up):
+        time_ms = get_time_millis()
         self.is_up = is_up
-        self.e_state.state(ENUM_UP_DN_STATES[0] if is_up else ENUM_UP_DN_STATES[1])
+        self.e_state.labels(name=self.name, url=self.url, method=self.method, server=self.instance_prefix).state(ENUM_UP_DN_STATES[0] if is_up else ENUM_UP_DN_STATES[1])
+        self.set_collect_time(get_time_millis() - time_ms)
         self.set_update_time()
         self.print_trigger_info()
 
@@ -118,17 +147,21 @@ class RestValueData(AbstractData):
         self.value = value
         self.type = result_type
         self.path = result_path
-        metric_name = self.get_metric_name('rest_value', name)
-        self.g_value = get_gauge_metric(metric_name, 'Remote REST API Value ' + name)
-        self.set_value(value)
+        self.g_value = get_gauge_metric('das_rest_value',
+                                        'Remote REST API [name, url, method, server] Value',
+                                        ['name', 'url', 'method', 'server'])
+        self.g_value.labels(name=name, url=url, method=method, server=self.instance_prefix)
+        self.set_data(value)
 
-    def set_value(self, value):
+    def set_data(self, value):
+        time_ms = get_time_millis()
         self.value = value
         try:
-            self.g_value.set(int(value))
+            self.g_value.labels(name=self.name, url=self.url, method=self.method, server=self.instance_prefix).set(int(value))
         except:
-            self.g_value.set(0)
+            self.g_value.labels(name=self.name, url=self.url, method=self.method, server=self.instance_prefix).set(0)
 
+        self.set_collect_time(get_time_millis() - time_ms)
         self.set_update_time()
         self.print_trigger_info()
 
@@ -142,17 +175,21 @@ class ShellValueData(AbstractData):
         self.command = command
         self.value = value
         self.args = args
-        metric_name = self.get_metric_name('shell_value', name)
-        self.g_value = get_gauge_metric(metric_name, 'Shell Value ' + name)
-        self.set_value(value)
+        self.g_value = get_gauge_metric('das_shell_value',
+                                        'Shell [name, command, server] Value',
+                                        ['name', 'command', 'server'])
+        self.g_value.labels(name=name, command=command, server=self.instance_prefix)
+        self.set_data(value)
 
-    def set_value(self, value):
+    def set_data(self, value):
+        time_ms = get_time_millis()
         self.value = value
         try:
-            self.g_value.set(int(value))
+            self.g_value.labels(name=self.name, command=self.command, server=self.instance_prefix).set(int(value))
         except:
-            self.g_value.set(0)
+            self.g_value.labels(name=self.name, command=self.command, server=self.instance_prefix).set(0)
 
+        self.set_collect_time(get_time_millis() - time_ms)
         self.set_update_time()
         self.print_trigger_info()
 
@@ -164,38 +201,44 @@ class IcmpData(AbstractData):
         self.ip = ip
         self.count = count
         self.is_up = is_up
-        metric_name = self.get_metric_name('host_available', name)
-        self.e_state = get_enum_metric(metric_name, 'Host availability', ENUM_UP_DN_STATES)
-        self.set_status(is_up)
+        self.e_state = get_enum_metric('das_host_available',
+                                       'Host [name, ip, server] availability',
+                                       ENUM_UP_DN_STATES, ['name', 'ip', 'server'])
+        self.e_state.labels(name=name, ip=ip, server=self.instance_prefix)
+        self.set_data(is_up)
 
-    def set_status(self, is_up):
+    def set_data(self, is_up):
+        time_ms = get_time_millis()
         self.is_up = is_up
-        self.e_state.state(ENUM_UP_DN_STATES[0] if is_up else ENUM_UP_DN_STATES[1])
+        self.e_state.labels(name=self.name, ip=self.ip, server=self.instance_prefix).state(ENUM_UP_DN_STATES[0] if is_up else ENUM_UP_DN_STATES[1])
+        self.set_collect_time(get_time_millis() - time_ms)
         self.set_update_time()
         self.print_trigger_info()
 
 
 class InterfaceData(AbstractData):
-    g_sent: Counter
-    g_receive: Counter
+    g_all: Counter
     def __init__(self, name, iface, interval, sent, receive, prefix=''):
         super().__init__(name, interval, prefix)
         self.iface = iface
         self.sent = sent
         self.receive = receive
-        sent_metric_name = self.get_metric_name('net_interface_sent_bytes', name)
-        self.g_sent = get_counter_metric(sent_metric_name, 'Network Interface bytes sent')
-        receive_metric_name = self.get_metric_name('net_interface_receive_bytes', name)
-        self.g_receive = get_counter_metric(receive_metric_name, 'Network Interface bytes receive')
+        self.g_all = get_counter_metric('das_net_interface_bytes',
+                                        'Network Interface [name, server, metric=[sent,receive]] bytes',
+                                        ['name', 'server', 'metric'])
+        self.g_all.labels(name=name, server=self.instance_prefix, metric='sent')
+        self.g_all.labels(name=name, server=self.instance_prefix, metric='receive')
         self.set_data(sent, receive)
 
     def set_data(self, sent, receive):
+        time_ms = get_time_millis()
         sent_delta = sent - self.sent
         recv_delta = receive - self.receive
         self.sent = sent
         self.receive = receive
-        self.g_sent.inc(sent_delta)
-        self.g_receive.inc(recv_delta)
+        self.g_all.labels(name=self.name, server=self.instance_prefix, metric='sent').inc(sent_delta)
+        self.g_all.labels(name=self.name, server=self.instance_prefix, metric='receive').inc(recv_delta)
+        self.set_collect_time(get_time_millis() - time_ms)
         self.set_update_time()
         self.print_trigger_info()
 
@@ -206,14 +249,18 @@ class UptimeData(AbstractData):
     def __init__(self, interval, prefix=''):
         super().__init__('uptime', interval, prefix)
         self.uptime = 0
-        metric_name = self.get_metric_name('exporter', self.name)
-        self.c_uptime = get_counter_metric(metric_name, 'Exporter Uptime in seconds')
+        self.c_uptime = get_counter_metric('das_exporter',
+                                           'Exporter Uptime for [server] in seconds',
+                                           ['server'])
+        self.c_uptime.labels(server=self.instance_prefix)
         self.set_data()
 
     def set_data(self):
+        time_ms = get_time_millis()
         uptime = int(time.time()) - self.START_TIME
-        self.c_uptime.inc(uptime - self.uptime)
+        self.c_uptime.labels(server=self.instance_prefix).inc(uptime - self.uptime)
         self.uptime = uptime
+        self.set_collect_time(get_time_millis() - time_ms)
         self.set_update_time()
         self.print_trigger_info()
 
@@ -232,23 +279,24 @@ class SystemData(AbstractData):
         self.set_data()
 
     def init_metrics(self):
-        uptime_metric_name = self.get_metric_name(self.name, 'uptime_seconds')
-        self.c_uptime = get_counter_metric(uptime_metric_name, 'System uptime')
-        cpu_metric_name = self.get_metric_name(self.name, 'cpu_percent')
-        self.g_cpu = get_gauge_metric(cpu_metric_name, 'CPU used percent')
-        mem_metric_name = self.get_metric_name(self.name, 'memory_percent')
-        self.g_memory = get_gauge_metric(mem_metric_name, 'Memory used percent')
-        chassis_temp_metric_name = self.get_metric_name(self.name, 'ChassisTemperature_current')
-        self.g_chassis_temp = get_gauge_metric(chassis_temp_metric_name, 'Current Chassis Temperature overall')
-        cpu_temp_metric_name = self.get_metric_name(self.name, 'CpuTemperature_current')
-        self.g_cpu_temp = get_gauge_metric(cpu_temp_metric_name, 'Current CPU Temperature overall')
+        self.c_uptime = get_counter_metric('das_uptime_seconds', 'System uptime on [server]', ['server'])
+        self.c_uptime.labels(server=self.instance_prefix)
+        self.g_cpu = get_gauge_metric('das_cpu_percent', 'CPU used percent on [server]', ['server'])
+        self.g_cpu.labels(server=self.instance_prefix)
+        self.g_memory = get_gauge_metric('das_memory_percent', 'Memory used percent on [server]', ['server'])
+        self.g_memory.labels(server=self.instance_prefix)
+        self.g_chassis_temp = get_gauge_metric('das_ChassisTemperature_current', 'Current Chassis Temperature overall on [server]', ['server'])
+        self.g_chassis_temp.labels(server=self.instance_prefix)
+        self.g_cpu_temp = get_gauge_metric('das_CpuTemperature_current', 'Current CPU Temperature overall on [server]', ['server'])
+        self.g_cpu_temp.labels(server=self.instance_prefix)
 
     def set_data(self):
+        time_ms = get_time_millis()
         uptime = int(time.time()) - self.BOOT_TIME
-        self.c_uptime.inc(uptime - self.uptime)
+        self.c_uptime.labels(server=self.instance_prefix).inc(uptime - self.uptime)
         self.uptime = uptime
         self.memory = psutil.virtual_memory().percent
-        self.g_memory.set(self.memory)
+        self.g_memory.labels(server=self.instance_prefix).set(self.memory)
         Thread(target=self.set_cpu_percent()).run()
 
         try:
@@ -272,20 +320,21 @@ class SystemData(AbstractData):
             else:
                 self.ch_temp = self.cpu_temp
 
-            self.g_chassis_temp.set(self.ch_temp)
-            self.g_cpu_temp.set(self.cpu_temp)
+            self.g_chassis_temp.labels(server=self.instance_prefix).set(self.ch_temp)
+            self.g_cpu_temp.labels(server=self.instance_prefix).set(self.cpu_temp)
         except:
             self.ch_temp = -500
             self.cpu_temp = -500
-            self.g_chassis_temp.set(self.ch_temp)
-            self.g_cpu_temp.set(self.cpu_temp)
+            self.g_chassis_temp.labels(server=self.instance_prefix).set(self.ch_temp)
+            self.g_cpu_temp.labels(server=self.instance_prefix).set(self.cpu_temp)
 
+        self.set_collect_time(get_time_millis() - time_ms)
         self.set_update_time()
         self.print_trigger_info()
 
     def set_cpu_percent(self):
         self.cpu = psutil.cpu_percent(1)
-        self.g_cpu.set(self.cpu)
+        self.g_cpu.labels(server=self.instance_prefix).set(self.cpu)
 
 
 if __name__ == '__main__':

metrics/MetricClasses.py

@@ -4,7 +4,6 @@ from abc import abstractmethod
 import time
 import platform
 import subprocess
-from linecache import cache
 
 import requests
 import psutil
@@ -15,6 +14,7 @@ from threading import Thread
 from metrics.DataStructures import DiskData, HealthData, IcmpData, ENUM_UP_DN_STATES, InterfaceData, UptimeData, \
     SystemData, RestValueData, ShellValueData
 
+
 class AbstractMetric:
     metric_key = ""
     config = {}
@@ -174,7 +174,7 @@ class HealthMetric(AbstractMetric):
     def proceed_metric(self):
         for d in self.data_array:
             if d.is_need_to_update():
-                thread = Thread(target=is_health_check, args=(d.url, d.timeout, d.method, d.user, d.password, d.headers, d.set_status))
+                thread = Thread(target=is_health_check, args=(d.url, d.timeout, d.method, d.user, d.password, d.headers, d.set_data))
                 thread.start()
 
     def print_debug_info(self):
@@ -193,7 +193,7 @@ class IcmpMetric(AbstractMetric):
     def proceed_metric(self):
         for d in self.data_array:
             if d.is_need_to_update():
-                thread = Thread(target=is_ping, args=(d.ip, d.count, d.set_status))
+                thread = Thread(target=is_ping, args=(d.ip, d.count, d.set_data))
                 thread.start()
 
     def print_debug_info(self):
@@ -244,7 +244,7 @@ class RestValueMetric(AbstractMetric):
         for d in self.data_array:
             if d.is_need_to_update():
                 thread = Thread(target=get_rest_value, args=(d.url, d.timeout, d.method, d.user, d.password, d.headers,
-                                                             d.set_value, d.type, d.path))
+                                                             d.set_data, d.type, d.path))
                 thread.start()
 
     def print_debug_info(self):
@@ -263,7 +263,7 @@ class ShellValueMetric(AbstractMetric):
     def proceed_metric(self):
         for d in self.data_array:
             if d.is_need_to_update():
-                thread = Thread(target=get_shell_value, args=(d.command, d.args, d.set_value))
+                thread = Thread(target=get_shell_value, args=(d.command, d.args, d.set_data))
                 thread.start()
 
     def print_debug_info(self):