scrapy-metrics-exporter 0.1.0

Real-time metrics exporter for Scrapy with automatic collection of all stats

Scrapy Metrics Exporter

Export Scrapy metrics in real-time with automatic collection of all Scrapy stats and custom metrics.

Current backend: Prometheus (OpenMetrics format)

Installation

pip install scrapy-metrics-exporter

Quick Start

Add to your Scrapy project's settings.py:

EXTENSIONS = {
    'scrapy_metrics_exporter.extension.MetricsExporter': 500,
}

That's it! Run your spider and metrics will be available at:

http://localhost:9410/metrics

Configuration

Optional settings (with defaults):

# Port for metrics server
METRICS_EXPORTER_PORT = 9410

# Host to bind to
METRICS_EXPORTER_HOST = '0.0.0.0'

# Stats update interval in seconds
METRICS_EXPORTER_UPDATE_INTERVAL = 30

Features

✨ Automatic Metrics Collection

• Exports all numeric Scrapy stats automatically
• No manual configuration needed for new metrics
• Custom stats created in spiders/pipelines are exported automatically

📊 Built-in Real-time Metrics

• Items scraped/dropped (Counters with spider label)
• HTTP requests by method (Counters with spider + method labels)
• HTTP responses by status (Counters with spider + status labels)
• Pending/active requests (Gauges)

🔧 Flexible & Extensible

• Works with any Scrapy middleware or pipeline
• Support for custom metrics via Scrapy stats API
• Easy to add Histograms/Summaries when needed

Available Metrics

Real-time Counters (via signals)

• scrapy_items_scraped_total{spider="..."} - Total items scraped
• scrapy_items_dropped_total{spider="..."} - Total items dropped
• scrapy_requests_total{spider="...", method="..."} - Total requests by method
• scrapy_responses_total{spider="...", status="..."} - Total responses by status

State Gauges

• scrapy_requests_pending{spider="..."} - Current pending requests
• scrapy_requests_active{spider="..."} - Current active requests

Automatic Scrapy Stats (examples)

• scrapy_memusage_max{spider="..."} - Maximum memory usage
• scrapy_downloader_request_bytes{spider="..."} - Total bytes downloaded
• scrapy_item_scraped_count{spider="..."} - Item count from Scrapy stats
• And all other numeric Scrapy stats automatically!

Custom Metrics

Any numeric stat you create in your spider or pipeline is automatically exported:

# In your spider or pipeline
self.crawler.stats.inc_value('custom/my_metric')
self.crawler.stats.set_value('custom/avg_price', 42.5)

Appears automatically as:

scrapy_custom_my_metric{spider="..."} 10.0
scrapy_custom_avg_price{spider="..."} 42.5

Query Examples (Prometheus)

# Total items scraped across all spiders
sum(scrapy_items_scraped_total)

# Items scraped by specific spider
scrapy_items_scraped_total{spider="my_spider"}

# Request rate per second
rate(scrapy_requests_total[5m])

# Top spiders by item count
topk(5, scrapy_items_scraped_total)

Future Backends

The exporter is designed to be backend-agnostic. While it currently uses Prometheus (OpenMetrics format), the architecture allows for easy addition of other backends:

• Prometheus (current) - OpenMetrics format via /metrics endpoint
• StatsD (planned) - Send metrics to StatsD/Graphite
• InfluxDB (planned) - Direct InfluxDB integration
• OpenTelemetry (planned) - OTLP protocol support
• Custom - Easy to implement your own backend

Architecture

The exporter uses a modular design:

• Collectors - Gather metrics from Scrapy (signals + stats)
• Registry - Store and manage metrics
• Backend - Format and expose metrics (currently Prometheus)
• Server - HTTP server for metrics endpoint (Twisted)

This separation makes it easy to add new backends without changing the core collection logic.

License

MIT