Support UTF-8 in metric creation, parsing, and exposition

part of #1013

Signed-off-by: Owen Williams <owen.williams@grafana.com>

Author: ywwg

prometheus_client/exposition.py

@@ -20,6 +20,7 @@
 from .openmetrics import exposition as openmetrics
 from .registry import CollectorRegistry, REGISTRY
 from .utils import floatToGoString
+from .validation import _is_valid_legacy_metric_name
 
 __all__ = (
     'CONTENT_TYPE_LATEST',
@@ -247,19 +248,26 @@ class TmpServer(ThreadingWSGIServer):
 def generate_latest(registry: CollectorRegistry = REGISTRY) -> bytes:
     """Returns the metrics from the registry in latest text format as a string."""
 
-    def sample_line(line):
-        if line.labels:
-            labelstr = '{{{0}}}'.format(','.join(
+    def sample_line(samples):
+        if samples.labels:
+            labelstr = '{0}'.format(','.join(
                 ['{}="{}"'.format(
-                    k, v.replace('\\', r'\\').replace('\n', r'\n').replace('"', r'\"'))
-                    for k, v in sorted(line.labels.items())]))
+                    openmetrics.escape_label_name(k), openmetrics._escape(v))
+                    for k, v in sorted(samples.labels.items())]))
         else:
             labelstr = ''
         timestamp = ''
-        if line.timestamp is not None:
+        if samples.timestamp is not None:
             # Convert to milliseconds.
-            timestamp = f' {int(float(line.timestamp) * 1000):d}'
-        return f'{line.name}{labelstr} {floatToGoString(line.value)}{timestamp}\n'
+            timestamp = f' {int(float(samples.timestamp) * 1000):d}'
+        if _is_valid_legacy_metric_name(samples.name):
+            if labelstr:
+                labelstr = '{{{0}}}'.format(labelstr)
+            return f'{samples.name}{labelstr} {floatToGoString(samples.value)}{timestamp}\n'
+        maybe_comma = ''
+        if labelstr:
+            maybe_comma = ','
+        return f'{{{openmetrics.escape_metric_name(samples.name)}{maybe_comma}{labelstr}}} {floatToGoString(samples.value)}{timestamp}\n'
 
     output = []
     for metric in registry.collect():
@@ -282,8 +290,8 @@ def sample_line(line):
                 mtype = 'untyped'
 
             output.append('# HELP {} {}\n'.format(
-                mname, metric.documentation.replace('\\', r'\\').replace('\n', r'\n')))
-            output.append(f'# TYPE {mname} {mtype}\n')
+                openmetrics.escape_metric_name(mname), metric.documentation.replace('\\', r'\\').replace('\n', r'\n')))
+            output.append(f'# TYPE {openmetrics.escape_metric_name(mname)} {mtype}\n')
 
             om_samples: Dict[str, List[str]] = {}
             for s in metric.samples:
@@ -299,9 +307,9 @@ def sample_line(line):
             raise
 
         for suffix, lines in sorted(om_samples.items()):
-            output.append('# HELP {}{} {}\n'.format(metric.name, suffix,
-                                                    metric.documentation.replace('\\', r'\\').replace('\n', r'\n')))
-            output.append(f'# TYPE {metric.name}{suffix} gauge\n')
+            output.append('# HELP {} {}\n'.format(openmetrics.escape_metric_name(metric.name + suffix),
+                                                  metric.documentation.replace('\\', r'\\').replace('\n', r'\n')))
+            output.append(f'# TYPE {openmetrics.escape_metric_name(metric.name + suffix)} gauge\n')
             output.extend(lines)
     return ''.join(output).encode('utf-8')
 

prometheus_client/metrics.py

@@ -10,19 +10,21 @@
 
 from . import values  # retain this import style for testability
 from .context_managers import ExceptionCounter, InprogressTracker, Timer
-from .metrics_core import (
-    Metric, METRIC_LABEL_NAME_RE, METRIC_NAME_RE,
-    RESERVED_METRIC_LABEL_NAME_RE,
-)
+from .metrics_core import Metric
 from .registry import Collector, CollectorRegistry, REGISTRY
 from .samples import Exemplar, Sample
 from .utils import floatToGoString, INF
+from .validation import (
+    _validate_exemplar, _validate_labelnames, _validate_metric_name,
+)
 
 T = TypeVar('T', bound='MetricWrapperBase')
 F = TypeVar("F", bound=Callable[..., Any])
 
 
 def _build_full_name(metric_type, name, namespace, subsystem, unit):
+    if not name:
+        raise ValueError('Metric name should not be empty')
     full_name = ''
     if namespace:
         full_name += namespace + '_'
@@ -38,31 +40,6 @@ def _build_full_name(metric_type, name, namespace, subsystem, unit):
     return full_name
 
 
-def _validate_labelname(l):
-    if not METRIC_LABEL_NAME_RE.match(l):
-        raise ValueError('Invalid label metric name: ' + l)
-    if RESERVED_METRIC_LABEL_NAME_RE.match(l):
-        raise ValueError('Reserved label metric name: ' + l)
-
-
-def _validate_labelnames(cls, labelnames):
-    labelnames = tuple(labelnames)
-    for l in labelnames:
-        _validate_labelname(l)
-        if l in cls._reserved_labelnames:
-            raise ValueError('Reserved label metric name: ' + l)
-    return labelnames
-
-
-def _validate_exemplar(exemplar):
-    runes = 0
-    for k, v in exemplar.items():
-        _validate_labelname(k)
-        runes += len(k)
-        runes += len(v)
-    if runes > 128:
-        raise ValueError('Exemplar labels have %d UTF-8 characters, exceeding the limit of 128')
-
 
 def _get_use_created() -> bool:
     return os.environ.get("PROMETHEUS_DISABLE_CREATED_SERIES", 'False').lower() not in ('true', '1', 't')
@@ -139,8 +116,7 @@ def __init__(self: T,
         self._documentation = documentation
         self._unit = unit
 
-        if not METRIC_NAME_RE.match(self._name):
-            raise ValueError('Invalid metric name: ' + self._name)
+        _validate_metric_name(self._name)
 
         if self._is_parent():
             # Prepare the fields needed for child metrics.

prometheus_client/metrics_core.py

@@ -1,15 +1,12 @@
-import re
 from typing import Dict, List, Optional, Sequence, Tuple, Union
 
 from .samples import Exemplar, NativeHistogram, Sample, Timestamp
+from .validation import _validate_metric_name
 
 METRIC_TYPES = (
     'counter', 'gauge', 'summary', 'histogram',
     'gaugehistogram', 'unknown', 'info', 'stateset',
 )
-METRIC_NAME_RE = re.compile(r'^[a-zA-Z_:][a-zA-Z0-9_:]*$')
-METRIC_LABEL_NAME_RE = re.compile(r'^[a-zA-Z_][a-zA-Z0-9_]*$')
-RESERVED_METRIC_LABEL_NAME_RE = re.compile(r'^__.*$')
 
 
 class Metric:
@@ -24,8 +21,7 @@ class Metric:
     def __init__(self, name: str, documentation: str, typ: str, unit: str = ''):
         if unit and not name.endswith("_" + unit):
             name += "_" + unit
-        if not METRIC_NAME_RE.match(name):
-            raise ValueError('Invalid metric name: ' + name)
+        _validate_metric_name(name)
         self.name: str = name
         self.documentation: str = documentation
         self.unit: str = unit

prometheus_client/openmetrics/exposition.py

@@ -2,6 +2,7 @@
 
 
 from ..utils import floatToGoString
+from ..validation import _is_valid_legacy_labelname, _is_valid_legacy_metric_name
 
 CONTENT_TYPE_LATEST = 'application/openmetrics-text; version=1.0.0; charset=utf-8'
 """Content type of the latest OpenMetrics text format"""
@@ -24,18 +25,27 @@ def generate_latest(registry):
         try:
             mname = metric.name
             output.append('# HELP {} {}\n'.format(
-                mname, metric.documentation.replace('\\', r'\\').replace('\n', r'\n').replace('"', r'\"')))
-            output.append(f'# TYPE {mname} {metric.type}\n')
+                escape_metric_name(mname), _escape(metric.documentation)))
+            output.append(f'# TYPE {escape_metric_name(mname)} {metric.type}\n')
             if metric.unit:
-                output.append(f'# UNIT {mname} {metric.unit}\n')
+                output.append(f'# UNIT {escape_metric_name(mname)} {metric.unit}\n')
             for s in metric.samples:
-                if s.labels:
-                    labelstr = '{{{0}}}'.format(','.join(
-                        ['{}="{}"'.format(
-                            k, v.replace('\\', r'\\').replace('\n', r'\n').replace('"', r'\"'))
-                            for k, v in sorted(s.labels.items())]))
+                if not _is_valid_legacy_metric_name(s.name):
+                    labelstr = escape_metric_name(s.name)
+                    if s.labels:
+                        labelstr += ', '
                 else:
                     labelstr = ''
+                
+                if s.labels:
+                    items = sorted(s.labels.items())
+                    labelstr += ','.join(
+                        ['{}="{}"'.format(
+                            escape_label_name(k), _escape(v))
+                            for k, v in items])
+                if labelstr:
+                    labelstr = "{" + labelstr + "}"
+                    
                 if s.exemplar:
                     if not _is_valid_exemplar_metric(metric, s):
                         raise ValueError(f"Metric {metric.name} has exemplars, but is not a histogram bucket or counter")
@@ -59,16 +69,47 @@ def generate_latest(registry):
                 timestamp = ''
                 if s.timestamp is not None:
                     timestamp = f' {s.timestamp}'
-                output.append('{}{} {}{}{}\n'.format(
-                    s.name,
-                    labelstr,
-                    floatToGoString(s.value),
-                    timestamp,
-                    exemplarstr,
-                ))
+                if _is_valid_legacy_metric_name(s.name):
+                    output.append('{}{} {}{}{}\n'.format(
+                        s.name,
+                        labelstr,
+                        floatToGoString(s.value),
+                        timestamp,
+                        exemplarstr,
+                    ))
+                else:
+                    output.append('{} {}{}{}\n'.format(
+                        labelstr,
+                        floatToGoString(s.value),
+                        timestamp,
+                        exemplarstr,
+                    ))
         except Exception as exception:
             exception.args = (exception.args or ('',)) + (metric,)
             raise
 
     output.append('# EOF\n')
     return ''.join(output).encode('utf-8')
+
+
+def escape_metric_name(s: str) -> str:
+    """Escapes the metric name and puts it in quotes iff the name does not
+    conform to the legacy Prometheus character set.
+    """
+    if _is_valid_legacy_metric_name(s):
+        return s
+    return '"{}"'.format(_escape(s))
+
+
+def escape_label_name(s: str) -> str:
+    """Escapes the label name and puts it in quotes iff the name does not
+    conform to the legacy Prometheus character set.
+    """
+    if _is_valid_legacy_labelname(s):
+        return s
+    return '"{}"'.format(_escape(s))
+
+
+def _escape(s: str) -> str:
+    """Performs backslash escaping on \, \n, and " characters."""
+    return s.replace('\\', r'\\').replace('\n', r'\n').replace('"', r'\"')