LinDB

LinDB is an open-source time-series database for numeric monitoring data, aimed at teams running IoT or IT operations systems of their own. It is most compelling when horizontal scale and multi-data-center operation matter more than sub-second precision or support for text records.

Overview

LinDB is designed for metrics from infrastructure, platforms, software services, applications, and IoT deployments. Its schema-free model organizes data as metrics, tags, and fields, which suits multidimensional monitoring without requiring a fixed schema in advance.

The maker says a single server could support more than one million write transactions per second. Treat that as a stated capability, not a universal workload guarantee. Computation and storage are separated, and ETCD is the sole external dependency, so deployment has a focused dependency footprint but still requires operators to manage ETCD.

Key features

• Scaling and resilience: Horizontal scaling, multiple replicas, and operation across multiple IDCs support deployments that need to grow or span data centers. LinDB is described as able to keep working with one replica alive, helping preserve service when other replicas are unavailable.
• Cross-IDC data: Writes can go to either IDC, with queries and aggregation across IDCs. This suits distributed monitoring setups that need a broader view than one data center provides.
• Automatic rollups: Minute-, hour-, and day-level rollups are supported after database creation. These built-in intervals can simplify access to coarser summaries, though they do not change the underlying time-accuracy limit.
• Ingestion and access: LinDB offers client connections, an HTTP REST API for inserts and queries, and an admin console. Documented ingestion protocols include FlatBuffers, Protocol Buffers, and InfluxDB Line Protocol; Go and Java client libraries are available, with other languages left to community contributions.
• Query and operations: LinQL uses SQL-like syntax for querying and for cluster or database operations. A version 1 REST API under /api/v1/* also covers insert, query, and state operations.
• Deployment: Standalone and cluster modes are documented, with binary deployment and ETCD as the external dependency. Installation guidance covers Linux, Windows, and macOS, and a Docker guide covers Linux.

The central constraint is data scope: LinDB handles numeric time-series data only, with time accuracy up to the second. It is a poor fit for text-based records or workloads that require finer timestamp precision. The introduction also cites self-governance and defenses against malicious users, but those claims do not remove the need to assess an operator-managed deployment.

Pricing

LinDB is free and open source under Apache 2.0. Its LinDB plan costs nothing, with no separate free trial. There are no paid tiers or seat and usage caps described for this plan; the tradeoff is that teams operate the software themselves rather than selecting a priced hosted plan.

Platforms

LinDB is documented for Linux, Windows, and macOS installation, with Docker deployment guidance for Linux. It also provides a web admin console, client connections, and an HTTP API. Self-hosted operation is central to its deployment model; the API and client interfaces let applications connect programmatically.

Who it's for

LinDB fits engineering teams building self-managed monitoring or IoT systems that collect numeric metrics across infrastructure and application layers. It is especially relevant where horizontal growth, replication, cross-IDC writes and queries, and automatic rollups are useful. Teams needing text records, sub-second timestamps, or a managed service should look elsewhere.

Pros and cons

• Pros: Horizontal scaling and replicas suit growing deployments, and the stated ability to operate with one replica alive adds resilience.
• Pros: Multi-IDC writes and queries let teams combine monitoring data across data centers.
• Pros: Automatic minute, hour, and day rollups provide standard summary intervals without requiring teams to add those rollups manually.
• Pros: Free Apache 2.0 software avoids a license charge and permits self-managed deployment.
• Cons: Numeric records only and second-level accuracy exclude text-heavy and finer-grained time-series use cases.
• Cons: ETCD remains an operational dependency, and the documented binary and cluster deployment model calls for infrastructure ownership.
• Cons: Go and Java are the documented client libraries; teams using other languages may need to contribute a client or use the HTTP interface.

Alternatives

For a broader comparison, browse Time-Series Databases.

• InfluxDB is a freemium option with a free plan and a pay-for-what-you-use Cloud Serverless data-in plan; consider it when that pricing model is a better fit.
• TDengine offers an open-source edition with clustering under the AGPL license, making it an alternative when that license and edition suit your deployment.
• QuestDB has a free Apache 2.0 open-source, self-managed single-instance plan with community support; choose it when that single-instance model fits.
• VictoriaMetrics offers open-source single-node and cluster versions; consider it when you want those versions as the basis of a deployment.
• Prometheus is a free Apache 2.0 monitoring and alerting toolkit; choose it when a toolkit rather than LinDB's database is the better match.
• Apache IoTDB is a free open-source time-series database with downloadable binaries and source.
• DolphinDB offers a perpetual-license Community Edition capped at two nodes, with two CPU cores and 8 GB RAM per node; that defined node limit may suit smaller deployments.
• Grafana k6 is a Linux-only freemium option whose free plan is limited to 500 virtual user hours per month; consider it when that allowance fits your needs.

Verdict

Choose LinDB for self-hosted numeric monitoring when horizontal scale, replication, rollups, and multi-IDC operation are priorities and second-level timestamps are sufficient. Its free Apache 2.0 license is a clear advantage, but teams handling text records or needing finer time accuracy should choose a different tool.