Merge pull request #2384 from larribas/2066-statsd-socket

Allow reporting StatsD metrics over UDS sockets

docs/source/settings.rst

@@ -420,7 +420,12 @@ environment variable ``PYTHONUNBUFFERED`` .
 
 **Default:** ``None``
 
-``host:port`` of the statsd server to log to.
+The address of the StatsD server to log to.
+
+Address is a string of the form:
+
+* ``unix://PATH`` : for a unix domain socket.
+* ``HOST:PORT`` : for a network address
 
 .. versionadded:: 19.1
 

gunicorn/config.py

@@ -514,15 +514,25 @@ def validate_chdir(val):
     return path
 
 
-def validate_hostport(val):
+def validate_statsd_address(val):
     val = validate_string(val)
     if val is None:
         return None
-    elements = val.split(":")
-    if len(elements) == 2:
-        return (elements[0], int(elements[1]))
-    else:
-        raise TypeError("Value must consist of: hostname:port")
+
+    # As of major release 20, util.parse_address would recognize unix:PORT
+    # as a UDS address, breaking backwards compatibility. We defend against
+    # that regression here (this is also unit-tested).
+    # Feel free to remove in the next major release.
+    unix_hostname_regression = re.match(r'^unix:(\d+)$', val)
+    if unix_hostname_regression:
+        return ('unix', int(unix_hostname_regression.group(1)))
+
+    try:
+        address = util.parse_address(val, default_port='8125')
+    except RuntimeError:
+        raise TypeError("Value must be one of ('host:port', 'unix://PATH')")
+
+    return address
 
 
 def validate_reload_engine(val):
@@ -1630,9 +1640,14 @@ class StatsdHost(Setting):
     cli = ["--statsd-host"]
     meta = "STATSD_ADDR"
     default = None
-    validator = validate_hostport
+    validator = validate_statsd_address
     desc = """\
-    ``host:port`` of the statsd server to log to.
+    The address of the StatsD server to log to.
+
+    Address is a string of the form:
+
+    * ``unix://PATH`` : for a unix domain socket.
+    * ``HOST:PORT`` : for a network address
 
     .. versionadded:: 19.1
     """

gunicorn/instrument/statsd.py

@@ -24,14 +24,17 @@ class Statsd(Logger):
     """statsD-based instrumentation, that passes as a logger
     """
     def __init__(self, cfg):
-        """host, port: statsD server
-        """
         Logger.__init__(self, cfg)
         self.prefix = sub(r"^(.+[^.]+)\.*$", "\\g<1>.", cfg.statsd_prefix)
+
+        if isinstance(cfg.statsd_host, str):
+            address_family = socket.AF_UNIX
+        else:
+            address_family = socket.AF_INET
+
         try:
-            host, port = cfg.statsd_host
-            self.sock = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
-            self.sock.connect((host, int(port)))
+            self.sock = socket.socket(address_family, socket.SOCK_DGRAM)
+            self.sock.connect(cfg.statsd_host)
         except Exception:
             self.sock = None
 

tests/test_config.py

@@ -318,6 +318,30 @@ def test_nworkers_changed():
     assert c.nworkers_changed(1, 2, 3) == 3
 
 
+def test_statsd_host():
+    c = config.Config()
+    assert c.statsd_host is None
+    c.set("statsd_host", "localhost")
+    assert c.statsd_host == ("localhost", 8125)
+    c.set("statsd_host", "statsd:7777")
+    assert c.statsd_host == ("statsd", 7777)
+    c.set("statsd_host", "unix:///path/to.sock")
+    assert c.statsd_host == "/path/to.sock"
+    pytest.raises(TypeError, c.set, "statsd_host", 666)
+    pytest.raises(TypeError, c.set, "statsd_host", "host:string")
+
+
+def test_statsd_host_with_unix_as_hostname():
+    # This is a regression test for major release 20. After this release
+    # we should consider modifying the behavior of util.parse_address to
+    # simplify gunicorn's code
+    c = config.Config()
+    c.set("statsd_host", "unix:7777")
+    assert c.statsd_host == ("unix", 7777)
+    c.set("statsd_host", "unix://some.socket")
+    assert c.statsd_host == "some.socket"
+
+
 def test_statsd_changes_logger():
     c = config.Config()
     assert c.logger_class == glogging.Logger

tests/test_statsd.py

@@ -59,6 +59,18 @@ def test_statsd_fail():
     logger.exception("No impact on logging")
 
 
+def test_statsd_host_initialization():
+    c = Config()
+    c.set('statsd_host', 'unix:test.sock')
+    logger = Statsd(c)
+    logger.info("Can be initialized and used with a UDS socket")
+
+    # Can be initialized and used with a UDP address
+    c.set('statsd_host', 'host:8080')
+    logger = Statsd(c)
+    logger.info("Can be initialized and used with a UDP socket")
+
+
 def test_dogstatsd_tags():
     c = Config()
     tags = 'yucatan,libertine:rhubarb'