Community Enterprise Observability Platform (CEOP)

100% OpenTelemetry Compliant • Built with DDD/CQRS • Production-Ready • Apache 2.0 Licensed

TelemetryFlow is an enterprise-grade, open-source observability platform that provides unified telemetry collection, storage, analysis, and visualization. It is 100% OpenTelemetry Protocol (OTLP) compliant, designed as an open-source alternative to commercial solutions like Datadog, New Relic, and Dynatrace.

OpenTelemetry Native

• 100% OTLP Compliance — Full support for metrics, logs, traces, and exemplars
• Zero Vendor Lock-in — Standard OpenTelemetry SDKs and collectors
• Dual Endpoint Support — Community v1 + Platform v2 on same collector

Enterprise Architecture

• Domain-Driven Design — 25+ bounded contexts with clear module isolation
• CQRS Implementation — Optimized read/write with 40+ command/query handlers
• Event-Driven — NATS + BullMQ hybrid messaging for real-time events
• Multi-Tenancy — Hierarchical isolation (Region → Organization → Workspace → Tenant)

Security First

• 5-Tier RBAC System — Granular role-based access control
• AWS-Style API Keys — Dual-key authentication (tfk-/tfs-) with Argon2id hashing
• MFA + SSO — TOTP, Google, GitHub, Azure AD, Okta, SAML, OIDC
• Complete Audit Trail — Every action logged to ClickHouse

graph TB
    subgraph SDKs["Language SDKs"]
        PYSDK["Python SDK"]
        GOSDK["Go SDK"]
    end

    subgraph Collection["Data Collection"]
        AGENT["TFO Agent v1.2.0<br/>39+ Integrations"]
        COLLECTOR["TFO Collector v1.2.1<br/>OCB Native"]
    end

    subgraph Platform["Platform Core"]
        MONO["Platform Monolith<br/>NestJS + Vue 3"]
        VIZ["TFO-Viz<br/>Standalone Dashboard"]
    end

    subgraph AI["AI Layer"]
        GOMCP["Go MCP Server"]
        PYMCP["Python MCP Server"]
    end

    SDKs -->|"OTLP"| Collection
    Collection -->|"OTLP v1/v2"| Platform
    Collection -->|"OTLP"| VIZ
    Platform -->|"MCP"| AI

    style SDKs fill:#e8f5e9,stroke:#2e7d32,color:#000
    style Collection fill:#e3f2fd,stroke:#1565c0,color:#000
    style Platform fill:#fff3e0,stroke:#e65100,color:#000
    style AI fill:#f3e5f5,stroke:#6a1b9a,color:#000

The TFO Agent consolidates multiple monitoring tools into a single Go binary:

graph TB
    subgraph Replaced["Replaces These Tools"]
        PROM["Prometheus"]
        KSM["kube-state-metrics"]
        NE["node-exporter"]
        FB["FluentBit"]
        CAD["cAdvisor"]
    end

    subgraph Agent["TFO Agent (Go 1.26, OTEL SDK v1.43.0)"]
        NE_MOD["Node Exporter<br/>CPU, Memory, DiskIO,<br/>Filesystem, Network"]
        K8S_MOD["Kubernetes<br/>Nodes, Pods, Deployments,<br/>Services, HPA, PDB"]
        CAD_MOD["cAdvisor<br/>Container CPU, Memory,<br/>Network, Filesystem"]
        DB_MOD["Databases<br/>MySQL, PostgreSQL, MongoDB,<br/>MSSQL, ClickHouse, Aurora,<br/>CockroachDB, TimescaleDB"]
        EBPF_MOD["eBPF<br/>Syscalls, Network,<br/>File I/O, Scheduler"]
    end

    Replaced -.->|"Consolidated"| Agent
    Agent -->|"OTLP"| PLATFORM["TFO Platform"]

    style Replaced fill:#ffebee,stroke:#c62828,color:#000
    style Agent fill:#e8f5e9,stroke:#2e7d32,color:#000

Key Capabilities:

• 9 database collectors (MySQL, PostgreSQL, MongoDB, MSSQL, ClickHouse, CockroachDB, Aurora, TimescaleDB, SQLite3)
• 28 eBPF kernel-level metrics across 7 categories
• 39+ third-party integrations (Cloud, APM, OSS Observability, Streaming, Network)
• Docker container monitoring (32 per-container metrics)
• Disk-backed buffer for offline resilience
• Cross-platform: Linux, macOS, Windows

Enterprise-grade OTLP collector built on OpenTelemetry Collector Builder (OCB) with 85+ community components and 4 custom TFO components:

Pipeline Architecture:

• Traces: tfootlp → k8sattributes → batch → tfo + spanmetrics + servicegraph
• Metrics: tfootlp → k8sattributes → transform → batch → tfo + prometheus
• Logs: tfootlp → k8sattributes → batch → tfo

Query Analytics (QAN) with native collectors for 9 databases:

• MySQL/MariaDB/Percona — InnoDB, replication, Galera, query analytics
• PostgreSQL — pg_stat_statements, activity, bgwriter, replication
• Amazon Aurora — 60+ CloudWatch metrics via AWS SDK
• MongoDB — Server status, replica set, sharding, profiler
• MSSQL — Wait stats, perf counters, query store, index usage
• ClickHouse, CockroachDB, TimescaleDB, SQLite3

459 total registry entries across 23 module codes, rendered by Vue composables → 13 chart types via RegistryGraphPanel (3 variants: default/mini/panel).

graph LR
    RAW["Raw<br/>TTL: 7-30d"]
    ONE_M["1m Agg<br/>TTL: 30-90d"]
    ONE_H["1h Agg<br/>TTL: 90-180d"]
    ONE_D["1d Agg<br/>TTL: 365d+"]

    RAW -->|"MV"| ONE_M -->|"MV"| ONE_H -->|"MV"| ONE_D

    style RAW fill:#ffebee,stroke:#c62828,color:#000
    style ONE_M fill:#fff3e0,stroke:#e65100,color:#000
    style ONE_H fill:#e3f2fd,stroke:#1565c0,color:#000
    style ONE_D fill:#e8f5e9,stroke:#2e7d32,color:#000

# Clone and setup
git clone https://github.com/telemetryflow/telemetryflow-platform.git
cd telemetryflow-platform

# Start infrastructure
docker-compose --profile core up -d

# Install, migrate, seed
pnpm install && pnpm db:migrate && pnpm db:seed

# Start development
pnpm dev

We welcome contributions! See individual repository CONTRIBUTING.md files for guidelines.

• License: Apache 2.0
• Built by: DevOpsCorner Indonesia
• Website: telemetryflow.id
• Docs: docs.telemetryflow.id