LinDB vs RRDtool
LinDB
lindb.ioRRDtool
oss.oetiker.chSide by side
| Figure | LinDB | RRDtool |
|---|---|---|
| Price from | Free | Free |
| Free plan | Yes | Yes |
| Free trial | — | — |
| Open source | — | — |
| Web | Yes | — |
| Windows | Yes | Yes |
| macOS | Yes | — |
| Linux | Yes | Yes |
| API | Yes | Yes |
| Self-hosted | Yes | Yes |
| Plans published | 1 | 1 |
| Spec score · Time-Series Databases | 8.7 (#5) | 7.8 (#20) |
| Deployment | The documentation describes standalone and cluster modes, and deployment as a binary with ETCD as its external dependency. | RRDtool is distributed as source and is also listed with binary distributions for platforms including Debian Linux, Solaris, HP-UX, and Cygwin Windows. |
| Performance | The maker says a single server could support more than one million write TPS. | The maker describes RRDtool as a high performance system. |
| Support | The product page links to GitHub channels for feature requests, bug reports, and questions. | The maker says it finances RRDtool work through customer-specific add-on features and support contracts. |
Plans
In their own words
LinDB
LinDB is an open-source distributed time-series database for IoT and IT operations monitoring. It stores and analyzes metrics across IaaS, PaaS, SaaS and application layers, using a schema-free model of metrics, tags and fields. LinQL provides SQL-like syntax for queries and cluster or database operations; data can be inserted and queried through an HTTP REST API, client connections and an admin console. Ingestion supports FlatBuffers, Protocol Buffers and InfluxDB protocols, with Go and Java client libraries listed. The system separates computation from storage and identifies ETCD as its only external dependency. It supports replicas, horizontal scaling, operation across multiple IDCs, and automatic minute, hour and day rollups. Deployment documentation covers standalone and cluster modes, with packages for Linux, Windows and macOS and a Docker guide for Linux. LinDB is free under Apache 2.0. Its data scope is limited to numeric time series with accuracy up to the second; it is not suited to text records or finer time precision.