UTF-8 support in metric and label names (#1255)

Adds UTF-8 support for metric and label names.

These changes are based on the work done on the Prometheus common
libraries [here](prometheus/common#537) and
[here](prometheus/common#570)

- The `prometheus-metrics-exposition-formats` module will use the new
quoting syntax `{"foo"}` iff the metric does not conform to the legacy
name format (`foo{}`)
- The `prometheus-metrics-model` module has a new flag
(`nameValidationScheme`) that determines if validation is done using the
legacy or the UTF-8 scheme. This flag can be set via a property in the
properties file.
- Scrapers can announce via content negotiation that they support UTF-8
names by adding `escaping=allow-utf-8` in the Accept header. In cases
where UTF-8 is not available, metric providers can be configured to
escape names in a few different ways: values (`U__` UTF value escaping
for perfect round-tripping), underscores (all invalid chars become `_`),
dots (dots become `_dot_`, `_` becomes `__`, all other values become
`___`). Escaping has a global default
(`PrometheusNaming.DEFAULT_ESCAPING_SCHEME`) or can also be specified in
Accept header with the `escaping=` term, which can be `allow-utf-8` (for
UTF-8-compatible), `underscores`, `dots`, or `values`.
This should still be a noop for existing configurations because scrapers
will not be passing the escaping key in the Accept header. Existing
functionality is maintained.
- The `prometheus-metrics-exporter-pushgateway` module will
[escape](https://github.com/prometheus/proposals/blob/main/proposals/2023-08-21-utf8.md#text-escaping)
UTF-8 grouping keys in the URL path used when pushing metrics (see
prometheus/pushgateway#689)

Work towards prometheus/prometheus#13095

---------

Signed-off-by: Federico Torres <federico.torres@grafana.com>
Signed-off-by: Gregor Zeitlinger <gregor.zeitlinger@grafana.com>
Co-authored-by: Gregor Zeitlinger <gregor.zeitlinger@grafana.com>

Author: fedetorres93

.github/super-linter.env

@@ -14,8 +14,6 @@ VALIDATE_GO_MODULES=false
 VALIDATE_HTML=false
 # done by checkstyle
 VALIDATE_JAVA=false
-# contradicting with prettier
-VALIDATE_JAVASCRIPT_STANDARD=false
 # we have many duplicate code in our codebase for demo purposes
 VALIDATE_JSCPD=false
 VALIDATE_PYTHON_PYLINT=false

benchmarks/src/main/java/io/prometheus/metrics/benchmarks/TextFormatUtilBenchmark.java

@@ -1,5 +1,6 @@
 package io.prometheus.metrics.benchmarks;
 
+import io.prometheus.metrics.config.EscapingScheme;
 import io.prometheus.metrics.expositionformats.ExpositionFormatWriter;
 import io.prometheus.metrics.expositionformats.OpenMetricsTextFormatWriter;
 import io.prometheus.metrics.expositionformats.PrometheusTextFormatWriter;
@@ -69,14 +70,15 @@ public OutputStream openMetricsWriteToByteArray(WriterState writerState) throws
     // avoid growing the array
     ByteArrayOutputStream byteArrayOutputStream = writerState.byteArrayOutputStream;
     byteArrayOutputStream.reset();
-    OPEN_METRICS_TEXT_FORMAT_WRITER.write(byteArrayOutputStream, SNAPSHOTS);
+    OPEN_METRICS_TEXT_FORMAT_WRITER.write(
+        byteArrayOutputStream, SNAPSHOTS, EscapingScheme.ALLOW_UTF8);
     return byteArrayOutputStream;
   }
 
   @Benchmark
   public OutputStream openMetricsWriteToNull() throws IOException {
     OutputStream nullOutputStream = NullOutputStream.INSTANCE;
-    OPEN_METRICS_TEXT_FORMAT_WRITER.write(nullOutputStream, SNAPSHOTS);
+    OPEN_METRICS_TEXT_FORMAT_WRITER.write(nullOutputStream, SNAPSHOTS, EscapingScheme.ALLOW_UTF8);
     return nullOutputStream;
   }
 
@@ -85,14 +87,15 @@ public OutputStream prometheusWriteToByteArray(WriterState writerState) throws I
     // avoid growing the array
     ByteArrayOutputStream byteArrayOutputStream = writerState.byteArrayOutputStream;
     byteArrayOutputStream.reset();
-    PROMETHEUS_TEXT_FORMAT_WRITER.write(byteArrayOutputStream, SNAPSHOTS);
+    PROMETHEUS_TEXT_FORMAT_WRITER.write(
+        byteArrayOutputStream, SNAPSHOTS, EscapingScheme.ALLOW_UTF8);
     return byteArrayOutputStream;
   }
 
   @Benchmark
   public OutputStream prometheusWriteToNull() throws IOException {
     OutputStream nullOutputStream = NullOutputStream.INSTANCE;
-    PROMETHEUS_TEXT_FORMAT_WRITER.write(nullOutputStream, SNAPSHOTS);
+    PROMETHEUS_TEXT_FORMAT_WRITER.write(nullOutputStream, SNAPSHOTS, EscapingScheme.ALLOW_UTF8);
     return nullOutputStream;
   }
 

docs/content/config/config.md

@@ -15,7 +15,7 @@ Future releases will add more options, like configuration via environment variab
 Example:
 
 ```properties
-io.prometheus.exporter.httpServer.port = 9401
+io.prometheus.exporter.httpServer.port=9401
 ```
 
 The property above changes the port for the
@@ -71,13 +71,13 @@ metric only by specifying the metric name. Example:
 Let's say you have a histogram named `latency_seconds`.
 
 ```properties
-io.prometheus.metrics.histogramClassicUpperBounds = 0.2, 0.4, 0.8, 1.0
+io.prometheus.metrics.histogramClassicUpperBounds=0.2, 0.4, 0.8, 1.0
 ```
 
 The line above sets histogram buckets for all histograms. However:
 
 ```properties
-io.prometheus.metrics.latency_seconds.histogramClassicUpperBounds = 0.2, 0.4, 0.8, 1.0
+io.prometheus.metrics.latency_seconds.histogramClassicUpperBounds=0.2, 0.4, 0.8, 1.0
 ```
 
 The line above sets histogram buckets only for the histogram named `latency_seconds`.
@@ -170,10 +170,15 @@ See Javadoc for details.
 
 <!-- editorconfig-checker-disable -->
 
-| Name                                       | Javadoc                                                                                                                                           | Note |
-| ------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------- | ---- |
-| io.prometheus.exporter.pushgateway.address | [PushGateway.Builder.address()](</client_java/api/io/prometheus/metrics/exporter/pushgateway/PushGateway.Builder.html#address(java.lang.String)>) |      |
-| io.prometheus.exporter.pushgateway.scheme  | [PushGateway.Builder.scheme()](</client_java/api/io/prometheus/metrics/exporter/pushgateway/PushGateway.Builder.html#scheme(java.lang.String)>)   |      |
-| io.prometheus.exporter.pushgateway.job     | [PushGateway.Builder.job()](</client_java/api/io/prometheus/metrics/exporter/pushgateway/PushGateway.Builder.html#job(java.lang.String)>)         |      |
+| Name                                              | Javadoc                                                                                                                                                                                    | Note |
+| ------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ---- |
+| io.prometheus.exporter.pushgateway.address        | [PushGateway.Builder.address()](</client_java/api/io/prometheus/metrics/exporter/pushgateway/PushGateway.Builder.html#address(java.lang.String)>)                                          |      |
+| io.prometheus.exporter.pushgateway.scheme         | [PushGateway.Builder.scheme()](</client_java/api/io/prometheus/metrics/exporter/pushgateway/PushGateway.Builder.html#scheme(java.lang.String)>)                                            |      |
+| io.prometheus.exporter.pushgateway.job            | [PushGateway.Builder.job()](</client_java/api/io/prometheus/metrics/exporter/pushgateway/PushGateway.Builder.html#job(java.lang.String)>)                                                  |      |
+| io.prometheus.exporter.pushgateway.escapingScheme | [PushGateway.Builder.escapingScheme()](</client_java/api/io/prometheus/metrics/exporter/pushgateway/PushGateway.Builder.html#escapingScheme(io.prometheus.metrics.config.EscapingScheme)>) | (1)  |
 
 <!-- editorconfig-checker-enable -->
+
+(1) Escaping scheme can be `allow-utf-8`, `underscores`, `dots`, or `values` as described in
+[escaping schemes](https://github.com/prometheus/docs/blob/main/docs/instrumenting/escaping_schemes.md#escaping-schemes) <!-- editorconfig-checker-disable-line -->
+and in the [Unicode documentation]({{< relref "../exporters/unicode.md" >}}).

docs/content/exporters/filter.md

@@ -1,6 +1,6 @@
 ---
 title: Filter
-weight: 2
+weight: 3
 ---
 
 All exporters support a `name[]` URL parameter for querying only specific metric names. Examples:

docs/content/exporters/httpserver.md

@@ -1,6 +1,6 @@
 ---
 title: HTTPServer
-weight: 3
+weight: 4
 ---
 
 The `HTTPServer` is a standalone server for exposing a metric endpoint. A minimal example

docs/content/exporters/pushgateway.md

@@ -1,6 +1,6 @@
 ---
 title: Pushgateway
-weight: 5
+weight: 6
 ---
 
 The [Prometheus Pushgateway](https://github.com/prometheus/pushgateway) exists to allow ephemeral

docs/content/exporters/servlet.md

@@ -1,6 +1,6 @@
 ---
 title: Servlet
-weight: 4
+weight: 5
 ---
 
 The

docs/content/exporters/spring.md

@@ -1,6 +1,6 @@
 ---
 title: Spring
-weight: 5
+weight: 7
 ---
 
 ## Alternative: Use Spring's Built-in Metrics Library

docs/content/exporters/unicode.md

@@ -0,0 +1,29 @@
+---
+title: Unicode
+weight: 2
+---
+
+The Prometheus Java client library allows all Unicode characters, that can be encoded as UTF-8.
+
+At scrape time, some characters are replaced based on the `encoding` header according
+to
+the [Escaping scheme](https://github.com/prometheus/docs/blob/main/docs/instrumenting/escaping_schemes.md). <!-- editorconfig-checker-disable-line -->
+
+For example, if you use the `underscores` escaping scheme, dots in metric and label names are
+replaced with underscores, so that the metric name `http.server.duration` becomes
+`http_server_duration`.
+
+Prometheus servers that do not support Unicode at all will not pass the `encoding` header, and the
+Prometheus Java client library will replace dots, as well as any character that is not in the legacy
+character set (`a-zA-Z0-9_:`), with underscores by default.
+
+When `escaping=allow-utf-8` is passed, add valid UTF-8 characters to the metric and label names
+without replacing them. This allows you to use dots in metric and label names, as well as
+other UTF-8 characters, without any replacements.
+
+## PushGateway
+
+When using the [Pushgateway](/exporters/pushgateway/), Unicode support has to be enabled
+explicitly by setting `io.prometheus.exporter.pushgateway.escapingScheme` to `allow-utf-8` in the
+Pushgateway configuration file - see
+[Pushgateway configuration]({{< relref "/config/config.md#exporter-pushgateway-properties" >}})

docs/content/otel/names.md

@@ -17,7 +17,7 @@ the Prometheus server as if you had exposed Prometheus metrics directly.
 
 The main steps when converting OpenTelemetry metric names to Prometheus metric names are:
 
-- Replace dots with underscores.
+- Escape illegal characters as described in [Unicode support]
 - If the metric has a unit, append the unit to the metric name, like `_seconds`.
 - If the metric type has a suffix, append it, like `_total` for counters.
 
@@ -29,14 +29,8 @@ OpenTelemetry's [Semantic Conventions for HTTP Metrics](https://opentelemetry.io
 say that if you instrument an HTTP server with OpenTelemetry, you must have a histogram named
 `http.server.duration`.
 
-Most names defined in semantic conventions use dots. In the Prometheus server, the dot is an illegal
-character (this might change in future versions of the Prometheus server).
+Most names defined in semantic conventions use dots.
+Dots in metric and label names are now supported in the Prometheus Java client library as
+described in [Unicode support].
 
-The Prometheus Java client library allows dots, so that you can use metric names and label names as
-defined in OpenTelemetry's semantic conventions.
-The dots will automatically be replaced with underscores if you expose metrics in Prometheus format,
-but you will see the original names with dots if you push your metrics in OpenTelemetry format.
-
-That way, you can use OTel-compliant metric and label names today when instrumenting your
-application with the Prometheus Java client, and you are prepared in case your monitoring backend
-adds features in the future that require OTel-compliant instrumentation.
+[Unicode support]: {{< relref "../exporters/unicode.md" >}}