README

Features

Tracing

• Using the official OpenTelemetry SDK
• Minimal auto-instrumentation for requests, console commands, consumers and doctrine
• Trace context propagation from incoming requests to consumers, outgoing http calls and databases (using sqlcommenter)
• Configurable blacklisting of requests by path to avoid useless traces, eg. /metrics or /_healthz
• Automatic log inclusion with configurable log level and channels

Metrics

• Using the Prometheus Client
• Minimal auto-instrumentation for common request, consumer and message metrics (see list of provided default metrics).
• Autoconfigurable metric providers, see below.

Logging

• Adds trace context to logs for correlation, with customizable keys.

Baggage

• Using the official OpenTelemetry SDK
• Baggage propagation from incoming requests to consumers and outgoing http calls

Health

• A simple endpoint to expose application health (default: /_healthz)
• Autoconfigurable healthcheck interface to add healthchecks to be made for global application health

Installation and configuration

composer require worldia/instrumentation-bundle <your-exporter>

You will aso need to install an exporter implementation and APCu is required by the prometheus exporter.

Add to ```bundles.php```:
```php
return [
    // Other bundles
    Instrumentation\InstrumentationBundle::class => ['all' => true],
];

Minimal configuration
See the complete configuration reference here or run bin/console config:dump-reference instrumentation.

// docker-compose.yaml

services:
  php:
    image: php:8.1
    environment:
      - OTEL_PHP_TRACES_PROCESSOR=batch
      - OTEL_TRACES_SAMPLER=parentbased_always_on
      - OTEL_TRACES_EXPORTER=otlp
      - OTEL_EXPORTER_OTLP_PROTOCOL=grpc
      - OTEL_EXPORTER_OTLP_ENDPOINT=http://jaeger:4317
  jaeger:
    image: jaegertracing/all-in-one:latest
    environment:
      COLLECTOR_OTLP_ENABLED: "true"

// instrumentation.yaml

instrumentation: ~

Usage

• Tracing
  • Simple tracing example
  • Simple tracing example using the static API
  • Add Urls to your traces in error messages
  • Use upstream request id as trace id
  • Customize operation (span) name for a message
  • Link strategy for a message
  • Propagating trace/baggage context in HTTP requests
  • Add request / response bodies as span attributes for HTTP requests
• Metrics
  • Adding a metric
  • Using Redis as storage adapter (Recommended)
• Logging
  • Customizing trace context log keys
• Health
  • Adding a healthcheck