OpenObserve supports multiple collection methods for metrics: *Metrics collection through Prometheus remote write

• Metrics collection via OTLP protocol using otelc-collector
• Custom metrics ingestion through HTTP API endpoints
• Infrastructure metrics collection through various integrations, including cloud provider services and the OpenObserve collector

OpenObserve handles various metric types:

• Standard metric types: counters, gauges, summary and histograms
• Prometheus-compatible metric types and aggregation methods
• System-level metrics (CPU, memory, disk)
• Application metrics (request rates, latencies, error rates)
• Custom metrics using any numeric data type with timestamps

The metric processing pipeline in OpenObserve is optimized for time-series data:

• Real-time processing of incoming metrics using downsampling or pipelines
• Storage in a columnar format optimized for time-series queries
• Support for various aggregation functions (rate calculations, moving averages, percentiles)
• Efficient indexing of labels and tags for fast filtering and grouping operations

OpenObserve provides a dual query interface for metrics:

• PromQL queries for compatibility with Prometheus workflows
• SQL for more complex analytics
• Support for time-based functions, mathematical operations, and advanced aggregations
• Cross-metric calculations
• Functions like rate and increase
• Complex metric expressions

The platform implements specific optimizations for high-cardinality metrics:

• OpenObserve stores data in parquet format and does not suffer from high cardinality issues
• Efficient label indexing and compression techniques
• Intelligent caching and query optimization to maintain performance

OpenObserve offers comprehensive visualization options for metrics:

• Time-series graphs, gauges, and heat maps
• Various panel types optimized for different metric visualizations
• Dynamic time ranges, zoom capabilities, and legend options
• Template variables for reusable dashboards
• Automatic refresh intervals
• Annotations for highlighting specific events in time
• Drill down and correlation between logs, metrics and traces

The platform provides native Prometheus compatibility:

• Support for prometheus remote write protocol to accept metrics from Prometheus agents
• PromQL support for queries
• Maintenance of label consistency with Prometheus metrics
• Support for all standard Prometheus metric types

OpenObserve provides comprehensive aggregation and alerting capabilities:

• Rolling aggregations and downsampling rules
• Custom metric transformations
• Threshold-based alerts and trend analysis
• Integration with various notification channels
• Detailed metric context in alerts