Quickwit

What is Quickwit?

Quickwit is an open source search engine for logs and traces that queries data directly on cloud object storage — S3, GCS, Azure Blob, MinIO, or Ceph — instead of expensive local disks. Written in Rust on top of the Tantivy library, it decouples compute from storage to deliver sub-second full-text search over massive observability datasets.

What is Quickwit best for?

Quickwit fits teams drowning in log and trace volume who want fast search without paying per-gigabyte SaaS bills or babysitting an Elasticsearch cluster. It shines for high-volume, lower-query-rate workloads — append-heavy observability data you search occasionally to troubleshoot — where storing everything on cheap object storage beats keeping it on hot SSDs.

What can Quickwit do?

• Search logs and traces directly on S3-compatible object storage with sub-second latency
• Expose an Elasticsearch-compatible search API, easing migration from an existing ES setup
• Ingest OpenTelemetry data natively and serve as a Jaeger-compatible trace backend
• Integrate with Grafana for dashboards and with Jaeger UI for trace exploration
• Support schemaless or strict-schema indexing, multi-tenancy, and retention policies
• Run stateless and horizontally scalable on Kubernetes via an official Helm chart
• Handle GDPR-style deletions with per-document delete tasks

Where does Quickwit fall short?

• Its future is uncertain. Datadog acquired Quickwit in January 2025 and the core team now builds inside Datadog. The project was relicensed from AGPL to Apache 2.0 and handed to the community, so ongoing development depends on community maintainers rather than a dedicated company.
• It’s search, not full observability. Quickwit covers logs and traces; metrics are still on the roadmap, so it isn’t a drop-in replacement for an all-in-one platform’s dashboards, alerting, and APM.
• Object-storage latency is a tradeoff. The cheap-storage architecture optimizes for high ingest and occasional search, not the high query rates of a live user-facing search backend — for that, a traditional Elasticsearch-style engine is a better fit.
• Highly available indexing needs a queue. Search is HA out of the box, but HA ingestion requires a distributed queue source such as Apache Kafka or Amazon Kinesis.

Is Quickwit free?

Yes — Quickwit is fully free and open source under the Apache 2.0 license, with no paid edition or feature gating. You self-host it and pay only for your own infrastructure, which is mostly cheap object storage; Quickwit estimates costs up to 10x lower than a comparable Elasticsearch deployment. There is no Quickwit-branded managed cloud.

What does Quickwit replace?

Quickwit is a self-hosted alternative to log and observability platforms like Splunk, Datadog logs, and Sumo Logic. It does the ingest-search-and-analyze job for logs and traces, but you run it on your own object storage and skip the volume-based licensing that makes those tools expensive at scale. It’s also positioned as a cheaper, lower-maintenance stand-in for a self-managed Elasticsearch logging stack.

FAQ

Is Quickwit open source? Yes. Quickwit is released under the permissive Apache 2.0 license. It was relicensed from AGPL to Apache 2.0 after Datadog’s 2025 acquisition to keep the community version going.

Can I self-host Quickwit for free? Yes. The software is free; you pay only for the object storage and compute you run it on. Because it searches data directly on S3-style storage, storage costs stay low.

Is Quickwit a good Splunk or Datadog alternative? For log and trace search at scale, yes — it delivers fast search on cheap storage without per-gigabyte pricing. But it doesn’t yet cover metrics, dashboards, and alerting, so it replaces the search layer rather than a full observability suite.

What do I need to run Quickwit? Object storage (S3, GCS, Azure Blob, MinIO, or Ceph) and a host or Kubernetes cluster. For highly available ingestion you also need a queue such as Kafka or Kinesis.