Quickwit

Capabilities

• Search directly on object storage Indexes live on S3-compatible storage and are searched there in sub-second time, so teams scale retention in cheap object capacity instead of provisioned cluster disks.[1]

• Decoupled stateless architecture Indexers and searchers hold no local state, which makes horizontal scaling a matter of adding compute nodes and keeps single-node, multi-node, on-premise, and cloud layouts on one design.[1]

• Schemaless log indexing JSON documents land in indexes without a predeclared mapping, so new log fields become searchable without migration steps during fast-moving incidents.[1]

• Elasticsearch and OpenSearch compatible API An api/v1/_elastic/ surface with bulk ingestion exists to ease migrations and tool integrations, letting existing shippers and scripts target Quickwit with limited rewrites.[1]

• Native observability integrations OpenTelemetry-native ingest plus Jaeger support and Grafana tutorials wire traces, logs, and metrics from collection to dashboard without custom glue.[1]

• Multi-tenancy with retention control Many indexes with partitioning, retention and lifecycle policies, and infrequent targeted deletion cover per-tenant data, expiry, and GDPR-style removal requests.[1]

• Full REST, query-language, and CLI surface A documented REST API, a dedicated query language, command-line options, and metrics endpoints give operators scriptable control over indexes, search, and monitoring.[1][2]

• Small-footprint self-hosted installs Linux and macOS binaries for x86-64 and aarch64, a Docker image, a Kubernetes Helm chart, and Glasskube install the same single binary on laptops and production hosts.[1]

Who will love it

• Log platform teams on object storage Teams already keeping data on S3-compatible storage will love searching it in place, because retention grows in cheap object capacity instead of cluster disks.[1]

• Tracing teams on OpenTelemetry Observability engineers shipping OTEL spans and Jaeger traces will love the native integrations, since collection, search, and Grafana dashboards connect without custom pipelines.[1]

• Cost-driven Elasticsearch migrators Teams paying for large Elasticsearch log clusters will love the compatible bulk API and stateless scaling, which make a like-for-like cost pilot concrete.[1][2]

Who should think twice

• No verified managed offering This review verified self-hosted installation only and could not confirm a managed offering after the January 2025 Datadog announcement, so teams without operations capacity should avoid it until hosting is settled.[1][2]

• Event retrieval, not relevance-tuned app search Quickwit optimizes append-only event retrieval rather than tuned product or site search, so teams needing relevance controls and typo tolerance should avoid assuming it matches engines built for that job.[1]

• Uncertain independent roadmap The Datadog acquisition leaves the independent roadmap genuinely uncertain, so organizations that need a settled multi-year vendor plan should avoid committing before that picture clears.[1]

Fit and compatibility