Vector is a free, Rust-built, open-source tool for collecting, transforming, and routing logs and metrics, maintained by Datadog. Not a vector database.
Vector is an open-source data pipeline built specifically for observability data: logs, metrics, and traces. It sits between the systems producing telemetry, such as application servers, containers, and cloud services, and the destinations that store or analyze that data, giving teams a single tool to collect, filter, transform, and route everything.
The project traces back to Timber.io, a startup that originally built developer-friendly logging products before shifting focus to build Vector as an open-source infrastructure tool. Its first commits date to 2018, and it was designed from the start in Rust for speed, memory safety, and minimal resource overhead compared to older, JVM-based or interpreted log shippers.
In February 2021, Datadog acquired Timber Technologies specifically to bring Vector's technology and team in-house. Vector has remained open source since the acquisition and continues to support routing data to many destinations beyond Datadog itself, and it is now maintained by Datadog's dedicated open-source engineering team.
It is important not to confuse Vector with a vector database. Vector.dev has no connection to AI embeddings, similarity search, or retrieval-augmented generation; it is strictly a logs-and-metrics pipeline tool, a distinct category from vector databases like Pinecone, Weaviate, or Milvus.
Vector supports dozens of sources for ingesting data, including Kubernetes logs, Kafka, Syslog, cloud provider log streams, and application-level log files, along with dozens of sinks for sending processed data onward to destinations like Datadog, Elasticsearch, Splunk, S3, Loki, Prometheus, and Kafka.
Its transform layer lets teams parse, filter, sample, aggregate, and enrich log and metric data in flight using the Vector Remap Language (VRL), reducing the volume and cost of data sent to expensive downstream storage.
Vector deploys as a single, dependency-free binary that can run as a lightweight agent on each host, a sidecar in a container, or a centralized aggregator, and its Rust foundation gives it high throughput with a comparatively small memory and CPU footprint versus older JVM-based shippers like Logstash.
Vector is entirely free and open source. There is no paid tier, seat-based pricing, or usage cap on the software itself, since it is a self-hosted pipeline component rather than a hosted SaaS product.
Organizations pay only for the compute resources needed to run Vector and, separately, for whatever destination platforms they route data into, such as Datadog itself, Elasticsearch, or a cloud storage bucket, each of which has its own independent pricing.
Vector is a free, open-source tool for collecting, transforming, and routing observability data such as logs, metrics, and traces between systems and destinations.
No. Despite the name, Vector is not a vector database or AI embeddings store. It is an observability data pipeline tool, unrelated to vector search technologies like Pinecone or Milvus.
Vector was created by Timber.io, with development starting in 2018, and has been maintained by Datadog's open-source engineering team since Datadog acquired Timber Technologies in February 2021.
Yes. Vector is fully open source and free to self-host, with no license fees for the software itself.
No. Vector is vendor-neutral and can route data to many destinations, including Elasticsearch, Splunk, S3, Prometheus, Loki, and Kafka, in addition to Datadog.
Vector is written in Rust, which gives it strong performance and memory safety compared with older, interpreted or JVM-based log shippers.
Vector competes directly with Fluent Bit, Logstash, and the OpenTelemetry Collector as a pipeline for logs and metrics, generally emphasizing higher throughput and a unified configuration model for both logs and metrics.