LinDB

LinDB is an open-source database for numeric time-series metrics, aimed at teams operating monitoring or IoT systems. It is a strong fit when horizontal growth and cross-data-center queries matter more than text support or sub-second timestamps. Its central trade-off is clear: broad deployment options and automatic rollups, but a deliberately narrow data model.

Overview

LinDB targets IT operations and IoT monitoring across infrastructure, platforms, software services, and application layers. Its schema-free model groups data into metrics, tags, and fields, which suits multidimensional metrics without requiring a fixed schema up front.

The maker says a single server could support more than one million write transactions per second. That makes LinDB worth considering for high-ingest workloads, though the stated figure is not a guarantee for every deployment. Computation and storage are separated, and ETCD is its only external dependency: a focused dependency footprint, but one that still requires operators to run ETCD.

Key features

• Scale and resilience: Horizontal scaling and multiple replicas suit deployments expected to grow or needing redundancy. LinDB is described as able to keep working with one replica alive, a useful continuity property when other replicas are unavailable.
• Multi-IDC operation: Writes can go to either IDC, with queries and aggregation across IDCs. That is valuable for monitoring distributed infrastructure in one view rather than treating each data center in isolation.
• Automatic rollups: Minute, hour, and day rollups are supported after database creation. These intervals can help teams work with metrics at multiple time scales without making rollups an extra setup step.
• Ingestion and queries: Data ingestion supports FlatBuffers, Protocol Buffers, and InfluxDB protocols; an HTTP write endpoint accepts InfluxDB Line Protocol. Go and Java client libraries are provided, and other language contributions are invited. The version 1 HTTP REST API under /api/v1/* supports insert, query, and state operations. LinQL uses SQL-like syntax for data queries and cluster or database operations, so SQL familiarity may help, though it is not a claim of general SQL compatibility.
• Deployment: Standalone and cluster modes are documented, with binary deployment and ETCD as the external dependency. Linux, Windows, and macOS installation guides, plus a Docker guide for Linux, give operators several ways to run it.

The main constraint is also decisive: LinDB handles numeric time-series data with accuracy only up to the second. It is unsuitable for text records or workloads that need finer timing, so it should not be treated as a general-purpose event store.

Pricing

LinDB costs nothing under the open-source Apache 2.0 plan. There is no paid tier or trial to weigh here; readers take on deployment and operation themselves. That makes it a straightforward option for teams able to manage the software, while those seeking a hosted service should compare alternatives.

Platforms

LinDB supports API access and is available for Linux, macOS, Windows, web, and self-hosted use. The documented installation paths favor teams comfortable operating infrastructure; the Linux Docker guide adds a container-based deployment route.

Who it's for

Choose LinDB for self-managed IT operations or IoT monitoring when metrics span multiple infrastructure layers, scaling and cross-IDC aggregation matter, and second-level numeric data is sufficient. Look elsewhere if the workload depends on text records, sub-second timestamps, or avoiding the responsibility of running a database and ETCD.

Pros and cons

Pros

• Horizontal scaling, multiple replicas, and multi-IDC queries support growing or geographically distributed monitoring deployments.
• Automatic minute, hour, and day rollups provide useful time resolutions without a separate rollup setup step.
• Multiple ingestion protocols, Go and Java clients, and an HTTP API offer several integration paths.
• The Apache 2.0 open-source plan has a price of $0, with no paid tier required to use the database.

Cons

• Numeric-only records and second-level accuracy exclude text data and finer-grained timing workloads.
• ETCD remains a required external dependency, and the self-hosted deployment model puts operation on the user.
• Go and Java are the named client libraries; teams using another language may need to contribute or build their own integration.

Alternatives

Time-Series Databases is a useful starting point for comparing this category. Consider InfluxDB if its freemium offering, free trial, or pay-for-what-you-use Cloud Serverless - Data In plan better suits your approach to hosting and cost. Prometheus is another free, open-source monitoring and alerting toolkit for readers prioritizing that toolkit focus. QuestDB offers a free Apache 2.0 self-managed single-instance plan with community support, a fit to compare when that deployment scope matches your needs. TDengine is worth comparing if its open-source TSDB-OSS clustering under the AGPL license is a better fit. VictoriaMetrics offers open-source single-node and cluster versions for readers comparing those deployment choices. Apache IoTDB is another free open-source time-series database with downloadable binaries and source. DolphinDB has a free perpetual Community Edition capped at two nodes, each with two CPU cores and 8 GB RAM, for readers whose needs fit those limits. Grafana k6 is a different option to compare for load testing, with its Free plan capped at 500 virtual user hours per month.

Verdict

LinDB is a sound choice for teams building self-hosted monitoring or IoT systems that need to scale numeric metrics across replicas and data centers. Its open-source plan, rollups, and cross-IDC aggregation are meaningful strengths; choose another database if text records, sub-second precision, or a managed operating model is essential.