feat: Add common namespace for metrics (#40)

Author: khvn26

README.md

@@ -63,19 +63,21 @@ The following default metrics are exposed:
 ##### Common metrics
 
 - `flagsmith_build_info`: Has the labels `version` and `ci_commit_sha`.
-- `http_server_request_duration_seconds`: Histogram labeled with `method`, `route`, and `response_status`.
-- `http_server_requests_total`: Counter labeled with `method`, `route`, and `response_status`.
-- `task_processor_enqueued_tasks_total`: Counter labeled with `task_identifier`.
+- `flagsmith_http_server_request_duration_seconds`: Histogram labeled with `method`, `route`, and `response_status`.
+- `flagsmith_http_server_requests_total`: Counter labeled with `method`, `route`, and `response_status`.
+- `flagsmith_task_processor_enqueued_tasks_total`: Counter labeled with `task_identifier`.
 
 ##### Task Processor metrics
 
-- `task_processor_finished_tasks_total`: Counter labeled with `task_identifier` and `result` (`"success"`, `"failure"`).
-- `task_processor_task_duration_seconds`: Histogram labeled with `task_identifier` and `result` (`"success"`, `"failure"`).
+- `flagsmith_task_processor_finished_tasks_total`: Counter labeled with `task_identifier` and `result` (`"success"`, `"failure"`).
+- `flagsmith_task_processor_task_duration_seconds`: Histogram labeled with `task_identifier` and `result` (`"success"`, `"failure"`).
 
 ##### Guidelines
 
 Try to come up with meaningful metrics to cover your feature with when developing it. Refer to [Prometheus best practices][1] when naming your metric and labels.
 
+As a reasonable default, Flagsmith metrics are expected to be namespaced with the `"flagsmith_"` prefix.
+
 Define your metrics in a `metrics.py` module of your Django application — see [example][2]. Contrary to Prometheus Python client examples and documentation, please name a metric variable exactly as your metric name.
 
 It's generally a good idea to allow users to define histogram buckets of their own. Flagsmith accepts a `PROMETHEUS_HISTOGRAM_BUCKETS` setting so users can customise their buckets. To honour the setting, use the `common.prometheus.Histogram` class when defining your histograms. When using `prometheus_client.Histogram` directly, please expose a dedicated setting like so:
@@ -84,8 +86,8 @@ It's generally a good idea to allow users to define histogram buckets of their o
 import prometheus_client
 from django.conf import settings
 
-distance_from_earth_au = prometheus.Histogram(
-    "distance_from_earth_au",
+flagsmith_distance_from_earth_au = prometheus.Histogram(
+    "flagsmith_distance_from_earth_au",
     "Distance from Earth in astronomical units",
     buckets=settings.DISTANCE_FROM_EARTH_AU_HISTOGRAM_BUCKETS,
 )

src/common/gunicorn/logging.py

@@ -60,10 +60,10 @@ def access(
             "method": environ.get("REQUEST_METHOD") or "",
             "response_status": resp.status_code,
         }
-        metrics.http_server_request_duration_seconds.labels(**labels).observe(
+        metrics.flagsmith_http_server_request_duration_seconds.labels(**labels).observe(
             duration_seconds
         )
-        metrics.http_server_requests_total.labels(**labels).inc()
+        metrics.flagsmith_http_server_requests_total.labels(**labels).inc()
 
 
 class GunicornJsonCapableLogger(PrometheusGunicornLogger):

src/common/gunicorn/metrics.py

@@ -2,13 +2,13 @@
 
 from common.prometheus import Histogram
 
-http_server_requests_total = prometheus_client.Counter(
-    "http_server_requests_total",
+flagsmith_http_server_requests_total = prometheus_client.Counter(
+    "flagsmith_http_server_requests_total",
     "Total number of HTTP requests",
     ["route", "method", "response_status"],
 )
-http_server_request_duration_seconds = Histogram(
-    "http_server_request_duration_seconds",
+flagsmith_http_server_request_duration_seconds = Histogram(
+    "flagsmith_http_server_request_duration_seconds",
     "HTTP request duration in seconds",
     ["route", "method", "response_status"],
 )

src/task_processor/decorators.py

@@ -88,7 +88,7 @@ def delay(
             self.run_in_thread(args=args, kwargs=kwargs)
         else:
             logger.debug("Creating task for function '%s'...", task_identifier)
-            metrics.task_processor_enqueued_tasks_total.labels(
+            metrics.flagsmith_task_processor_enqueued_tasks_total.labels(
                 task_identifier=task_identifier
             ).inc()
             try:

src/task_processor/metrics.py

@@ -3,20 +3,20 @@
 
 from common.prometheus import Histogram
 
-task_processor_enqueued_tasks_total = prometheus_client.Counter(
-    "task_processor_enqueued_tasks_total",
+flagsmith_task_processor_enqueued_tasks_total = prometheus_client.Counter(
+    "flagsmith_task_processor_enqueued_tasks_total",
     "Total number of enqueued tasks",
     ["task_identifier"],
 )
 
 if settings.TASK_PROCESSOR_MODE:
-    task_processor_finished_tasks_total = prometheus_client.Counter(
-        "task_processor_finished_tasks_total",
+    flagsmith_task_processor_finished_tasks_total = prometheus_client.Counter(
+        "flagsmith_task_processor_finished_tasks_total",
         "Total number of finished tasks",
         ["task_identifier", "result"],
     )
-    task_processor_task_duration_seconds = Histogram(
-        "task_processor_task_duration_seconds",
+    flagsmith_task_processor_task_duration_seconds = Histogram(
+        "flagsmith_task_processor_task_duration_seconds",
         "Task processor task duration in seconds",
         ["task_identifier", "result"],
     )

src/task_processor/processor.py

@@ -109,7 +109,7 @@ def _run_task(
     )
 
     ctx = ExitStack()
-    timer = metrics.task_processor_task_duration_seconds.time()
+    timer = metrics.flagsmith_task_processor_task_duration_seconds.time()
     ctx.enter_context(timer)
 
     task_identifier = task.task_identifier
@@ -158,7 +158,7 @@ def _run_task(
     )  # type: ignore[no-untyped-call]
     ctx.close()
 
-    metrics.task_processor_finished_tasks_total.labels(
+    metrics.flagsmith_task_processor_finished_tasks_total.labels(
         task_identifier=task_identifier,
         result=result_label_value,
     ).inc()

tests/unit/common/gunicorn/test_logging.py

@@ -99,12 +99,12 @@ def test_gunicorn_prometheus_gunicorn_logger__expected_metrics(
 
     # Then
     assert_metric(
-        name="http_server_requests_total",
+        name="flagsmith_http_server_requests_total",
         value=1.0,
         labels={"method": "GET", "route": "^health", "response_status": "200"},
     )
     assert_metric(
-        name="http_server_request_duration_seconds_sum",
+        name="flagsmith_http_server_request_duration_seconds_sum",
         value=0.101,
         labels={"method": "GET", "route": "^health", "response_status": "200"},
     )

tests/unit/task_processor/test_unit_task_processor_decorators.py

@@ -220,7 +220,7 @@ def my_function(*args: typing.Any, **kwargs: typing.Any) -> None:
 
     # Then
     assert_metric(
-        name="task_processor_enqueued_tasks_total",
+        name="flagsmith_task_processor_enqueued_tasks_total",
         value=1.0,
         labels={"task_identifier": "test_unit_task_processor_decorators.my_function"},
     )

tests/unit/task_processor/test_unit_task_processor_processor.py

@@ -603,31 +603,31 @@ def test_run_tasks__expected_metrics(
 
     # Then
     assert_metric(
-        name="task_processor_finished_tasks_total",
+        name="flagsmith_task_processor_finished_tasks_total",
         value=1.0,
         labels={
             "task_identifier": dummy_task_identifier,
             "result": "success",
         },
     )
     assert_metric(
-        name="task_processor_finished_tasks_total",
+        name="flagsmith_task_processor_finished_tasks_total",
         value=1.0,
         labels={
             "task_identifier": raise_exception_task_identifier,
             "result": "failure",
         },
     )
     assert_metric(
-        name="task_processor_task_duration_seconds",
+        name="flagsmith_task_processor_task_duration_seconds",
         value=mocker.ANY,
         labels={
             "task_identifier": dummy_task_identifier,
             "result": "success",
         },
     )
     assert_metric(
-        name="task_processor_task_duration_seconds",
+        name="flagsmith_task_processor_task_duration_seconds",
         value=mocker.ANY,
         labels={
             "task_identifier": raise_exception_task_identifier,