Skip to content

High-performance log analytics and observability platform built on Apache Arrow and DuckDB. Cost-effective alternative to Datadog with enterprise features.

License

Notifications You must be signed in to change notification settings

realm9-platform/ro9-observability

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 

Repository files navigation

RO9 Observability Platform

Next-Generation Enterprise Observability Designed for Cost-Conscious Teams

Performance Latency Compression Cost

Overview

RO9 is a next-generation observability platform engineered from first principles to deliver enterprise-scale log management at a fraction of traditional costs. By combining cutting-edge columnar storage, intelligent indexing, and multi-tier architecture, RO9 is designed to achieve performance metrics that can compete with industry leaders while dramatically reducing infrastructure costs.

Why We Built RO9

After spending $60,000/year on Datadog for just 10GB/day of logs, we realized the observability industry was fundamentally broken. We built RO9 to prove that enterprise observability doesn't have to bankrupt your infrastructure budget.

Performance Design Goals

Target Performance Characteristics

Based on our architecture and initial testing, RO9 is designed to achieve the following ballpark metrics. Actual performance will vary based on workload characteristics, hardware configuration, and deployment environment.

Metric RO9 Target Industry Average*
Ingestion Rate Up to 200K/sec 30-50K/sec
Query Latency (P50) Sub-10ms goal 100-200ms
Query Latency (P99) Sub-50ms goal 800-1500ms
Compression Ratio 15-25:1 estimated 5-8:1
Storage Cost/TB ~$5-10 $50-120
Monthly Cost (10GB/day) From $75/month $2,000-5,000

*Industry averages based on publicly available information from vendor documentation and third-party benchmarks. Your results may vary.

Architecture Benefits

Our multi-tier storage design and columnar format provide several advantages:

  • Ingestion: Arrow IPC enables high-throughput data streaming with minimal overhead
  • Storage: Parquet columnar format with aggressive compression strategies
  • Queries: DuckDB vectorized execution for analytical workloads
  • Costs: Automatic tiering to S3/Glacier dramatically reduces storage expenses

Technical Architecture

Multi-Tier Storage Design

┌─────────────────────────────────────────────────────────┐
│                   Ingestion Layer                        │
│         (High-throughput via Arrow IPC)                  │
├─────────────────────────────────────────────────────────┤
│                    Hot Tier (Redis)                      │
│              Last 15 minutes - Zero latency              │
├─────────────────────────────────────────────────────────┤
│                   Warm Tier (NVMe SSD)                   │
│           Last 24 hours - Parquet format                 │
├─────────────────────────────────────────────────────────┤
│                   Cold Tier (S3/Blob)                    │
│         Historical data - Compressed Parquet             │
├─────────────────────────────────────────────────────────┤
│                   Archive Tier (Glacier)                 │
│              Long-term retention - 99% cost reduction    │
├─────────────────────────────────────────────────────────┤
│                 Query Engine (DuckDB)                    │
│         Distributed SQL with automatic tier routing      │
└─────────────────────────────────────────────────────────┘

Core Technologies

Apache Arrow IPC

  • Zero-copy data transfer between services
  • Columnar memory format for 10x compression
  • SIMD optimizations for parallel processing
  • Language-agnostic data representation

DuckDB Analytics Engine

  • Vectorized query execution for massive parallelism
  • Columnar storage with aggressive compression
  • SQL interface for familiar querying
  • Embedded design for minimal overhead

Bloom Filter Index

  • Probabilistic indexing for instant filtering
  • 1-bit per event memory footprint
  • Sub-millisecond lookups across billions of events
  • False positive rate < 0.1%

Multi-Tier Storage

  • Hot (Redis): Real-time data, instant access
  • Warm (NVMe): Recent data, <10ms access
  • Cold (S3): Historical data, <100ms access
  • Archive (Glacier): Compliance retention, <1hr access

Key Innovations

1. Intelligent Data Tiering

RO9 automatically moves data between storage tiers based on access patterns and age:

  • Hot Tier: In-memory for real-time data (recent logs)
  • Warm Tier: NVMe SSD for frequently accessed data (last 24 hours)
  • Cold Tier: S3 Standard for historical data (recent weeks)
  • Archive Tier: Glacier for long-term retention (compliance)

2. Adaptive Compression

  • Dictionary encoding for repeated strings (95% reduction)
  • Delta encoding for timestamps (80% reduction)
  • Run-length encoding for sparse data (90% reduction)
  • Zstandard compression for final storage (60% reduction)

3. Smart Query Routing

RO9 automatically routes queries to the most appropriate storage tier based on the time range and data characteristics, optimizing for both performance and cost.

4. Correlation Engine

  • Automatic trace correlation across logs, metrics, traces
  • Service dependency mapping without manual configuration
  • Anomaly detection using statistical analysis
  • Root cause analysis with ML-powered insights

Use Cases

Log Analytics

  • Ingest millions of events per second
  • Search across years of data in milliseconds
  • Complex aggregations and analytics
  • Real-time alerting and monitoring

Application Performance Monitoring

  • Distributed tracing with automatic correlation
  • Service mesh observability
  • Database query analysis
  • API performance tracking

Security Analytics

  • Real-time threat detection
  • Compliance audit logging
  • User behavior analytics
  • Forensic investigation

Business Intelligence

  • Customer journey tracking
  • Revenue analytics
  • Operational metrics
  • Custom dashboards

Getting Started

Quick Deploy (Helm)

# Deploy RO9 Observability
helm install ro9 oci://public.ecr.aws/m0k6f4y3/realm9/ro9-observability \
  --version 1.71.0 \
  --namespace ro9-system \
  --create-namespace \
  --set ingress.host=ro9.your-domain.com

Production Deployment (Kubernetes)

RO9 can be deployed on Kubernetes using Helm charts. The platform supports:

  • High Availability: Multiple replicas with automatic failover
  • Scalability: Horizontal scaling based on ingestion rate
  • Storage: Integration with cloud storage providers (S3, Azure Blob, GCS)
  • Monitoring: Built-in health checks and metrics endpoints

Integration

RO9 provides SDKs for popular programming languages (Node.js, Python, Go, Java) and supports standard protocols:

  • OpenTelemetry: Native OTLP ingestion for logs, metrics, and traces
  • REST API: Simple HTTP API for custom integrations
  • SDKs: Type-safe client libraries with automatic retry and batching

Cost Analysis

Estimated Monthly Costs (10GB/day workload)

Based on our multi-tier architecture and S3 storage economics, we estimate the following cost structure. Actual costs will vary based on usage patterns, retention policies, and cloud provider pricing.

Provider Total (est.) Notes
RO9 ~$75-150 Estimated based on S3 storage + compute
Datadog $5,000+ Based on published pricing
New Relic $4,500+ Based on published pricing
Splunk $6,000+ Based on published pricing
Elastic Cloud $2,000+ Based on published pricing

Cost Advantages

Our architecture aims to reduce costs through:

  • Multi-tier storage: Hot data in Redis, warm on NVMe, cold on S3
  • Aggressive compression: Parquet format with columnar compression
  • Automatic archival: Move old data to Glacier for long-term retention
  • Efficient indexing: Bloom filters reduce storage overhead

Integrations

Data Sources

  • Applications: OpenTelemetry, Jaeger, Zipkin
  • Infrastructure: Kubernetes, Docker, systemd
  • Cloud: AWS CloudWatch, Azure Monitor, GCP Logging
  • Custom: REST API, gRPC, Kafka, AMQP

Visualization

  • Native: Built-in dashboards and explorer
  • Grafana: Full Grafana datasource plugin
  • API: RESTful API for custom integrations

Alerting

  • Channels: Email, Slack, PagerDuty, Webhook
  • Rules: Threshold, anomaly, pattern matching
  • ML-Powered: Predictive alerts before issues occur

Security & Compliance

Data Protection

  • Encryption at rest: AES-256-GCM
  • Encryption in transit: TLS 1.3
  • Data masking: PII automatic detection and masking
  • Access control: RBAC with fine-grained permissions

Compliance

  • GDPR: Right to erasure, data portability
  • HIPAA: Audit logging, encryption standards
  • SOC 2: Access controls, monitoring
  • ISO 27001: Information security management

Architecture Deep Dive

Ingestion Pipeline

[Log Sources] → [Collectors] → [Arrow IPC] → [Stream Processor] → [Router]
                                    ↓              ↓                 ↓
                               [Compression]  [Enrichment]    [Tier Manager]

Query Execution

[SQL Query] → [Parser] → [Planner] → [Optimizer] → [Executor]
                ↓           ↓            ↓            ↓
           [Validation] [Tier Route] [Pushdown]  [DuckDB Engine]

Storage Hierarchy

Level 0: Redis (Real-time) - 15 minutes
Level 1: NVMe SSD (Hot) - 24 hours
Level 2: S3 Standard (Warm) - 30 days
Level 3: S3 IA (Cool) - 90 days
Level 4: Glacier (Archive) - 7 years

Performance Optimization

Tips for Maximum Performance

  1. Use structured logging - JSON format preferred
  2. Enable field indexing for frequently searched fields
  3. Implement sampling for high-volume, low-value logs
  4. Use batch ingestion for better throughput
  5. Optimize query patterns - filter before aggregating

Configuration

RO9 provides flexible configuration options for:

  • Ingestion: Batch size, compression algorithms, parallelism
  • Storage: Retention policies, tier thresholds, compression strategies
  • Query: Cache settings, memory limits, query timeouts
  • Security: Encryption, access control, audit logging

All settings can be adjusted based on your workload characteristics and performance requirements.

Roadmap

Q1 2025

  • Apache Arrow IPC implementation
  • DuckDB query engine
  • Multi-tier storage
  • Distributed query execution

Q2 2025

  • Machine learning anomaly detection
  • Predictive alerting
  • Automatic capacity planning
  • Cost optimization advisor

Q3 2025

  • Edge computing support
  • Federated queries
  • Custom retention policies
  • Data lineage tracking

Community

Support

  • Community: GitHub Issues
  • Professional: support@realm9.app
  • Enterprise: 24/7 support with SLA

License

Copyright © 2025 Realm9. All rights reserved.

RO9 is available under a dual license:

  • Community Edition: Apache 2.0 License
  • Enterprise Edition: Commercial License with support

RO9 - Observability Without the Observatory Price Tag

Part of the Realm9 Platform | Benchmarks | Architecture

About

High-performance log analytics and observability platform built on Apache Arrow and DuckDB. Cost-effective alternative to Datadog with enterprise features.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 3

  •  
  •  
  •