The company behind InfluxDB - the open source database built for the one dimension every dataset shares: time.
Nobody wakes up wanting a time series database. They want to know why the dashboard spiked at 3 a.m., how a factory robot behaved last Tuesday, or whether a trade cleared in the millisecond it was supposed to. Every one of those questions has the same shape - a value, and a timestamp. InfluxData built its business on that single, universal dimension.
The company makes InfluxDB, an open source database purpose-built to ingest and query timestamped data: the readings from IoT sensors, the metrics from servers and containers, the ticks from financial markets, the telemetry from vehicles and spacecraft. Where a general-purpose database strains under millions of writes per second and an ever-growing set of unique tags, InfluxDB is engineered for exactly that load.
Its origin is a lesson in paying attention to your own tooling. InfluxData began in 2012 as Errplane, a monitoring startup that went through Y Combinator. The monitoring product never took off - but the internal time series engine the team built to measure it did. The company pivoted, renamed itself InfluxData in 2015, and turned that tool into a category.
Today InfluxDB is consistently ranked the most popular time series database in the world, with more than a million active instances running and over 1,900 organizations paying for the commercial platform. It is used inside Tesla's manufacturing lines, PayPal's machine-learning pipelines, and Nordstrom's storefront monitoring.
"On a mission to help developers and organizations store and analyze real-time data - so they can build monitoring, analytics, and IoT applications faster and at scale."
The bet has always been narrow and deliberate: do one thing - time - better than a general database ever will, give the engine away to developers, and sell the operations to the enterprises that depend on it.
Time series workloads break the assumptions of ordinary databases. Writes arrive constantly and in enormous volume, queries are almost always scoped to a window of time, and the number of unique series - each sensor, each host, each tag combination - can explode into the billions. That last problem, "high cardinality," was for years the ceiling that time series databases hit and fell off.
Built to absorb millions of data points per second from fleets of devices and services without dropping a beat.
InfluxDB 3's columnar design removes the cardinality ceiling that throttled earlier time series engines.
Engineered for fast reads on recent data, with SQL and InfluxQL as first-class query languages.
Data lands as compressed Apache Parquet on inexpensive object storage, decoupling compute from cost.
Rewriting a database from scratch is usually how a company loses years and momentum. InfluxData did it anyway. Earlier versions of InfluxDB were written in Go; for version 3, the team rebuilt the core in Rust and re-founded it on an open-source stack the industry now calls FDAP - Flight, DataFusion, Apache Arrow, and Parquet.
The point was not novelty. Apache Arrow gives InfluxDB a fast columnar memory format; DataFusion supplies a modern SQL query engine; Parquet provides compact, portable storage; and Arrow Flight moves data efficiently over the wire. Because the data already lives in the formats analytics and AI tools expect, InfluxDB slots into data lakes and machine-learning pipelines without a translation layer.
InfluxData is itself a leading contributor to Arrow and DataFusion, which now underpin other databases too. That is the difference the rebuild bought: not a feature, but a foundation competitors would have to match rather than out-market.
"InfluxDB 3 supports millions of writes per second, billions of series with no cardinality limits, and sub-10ms query response times."
Against the alternatives. In monitoring, the open-source alternatives are Prometheus and Grafana; among databases, TimescaleDB, QuestDB, VictoriaMetrics, and Amazon Timestream compete for the same workloads. Prometheus excels at metrics collection and alerting; TimescaleDB brings a relational, PostgreSQL-native model. InfluxDB's pitch is a purpose-built platform - ingestion, storage, and query in one - that scales from a laptop to a multi-cloud cluster.
Its companion collector, Telegraf, ships with more than 300 plugins and has become a de facto standard for pulling metrics out of almost anything.
The InfluxDB 3 line runs from a free open-source binary a developer can start in minutes to fully managed, single-tenant cloud clusters. The model is deliberate: adopt for free, grow into the paid tiers as the workload does.
High-speed engine for recent time series data, built in Rust for real-time apps.
Adds high availability, enhanced security, and scale for production.
Fully managed, single-tenant service for larger production workloads.
Usage-based managed service for smaller or variable workloads.
Elastic InfluxDB 3 for on-premises and private cloud deployments.
Plugin-driven agent with 300+ integrations for collecting metrics and events.
Web UI for querying and visualization, now with an "Ask AI" natural-language beta.
InfluxData runs an open-core model. The database engine and the Telegraf collector are free and open source, which drives the wide developer adoption behind those million-plus instances. Money comes from the parts organizations do not want to run themselves - managed cloud (Serverless and Cloud Dedicated, billed by usage), self-managed Enterprise and Clustered licenses, and support subscriptions. A million free instances exist so that a couple thousand can pay.
More than 1,900 paying customers span DevOps and observability teams, industrial IoT, financial services, and consumer applications. A sample of the names that run InfluxDB:
InfluxData has raised steadily since its first institutional round in 2014, culminating in an $81M Series E (equity plus debt) in early 2023 - capital that helped fund the multi-year rebuild into InfluxDB 3.
| Round | Amount | Date | Lead Investors |
|---|---|---|---|
| Series A | $8.1M | Nov 2014 | Mayfield Fund, Trinity Ventures |
| Series B | $16M | Sep 2016 | Battery Ventures |
| Series C | $35M | Feb 2018 | Sapphire Ventures |
| Series D | $60M | 2019 | Norwest Venture Partners |
| Series E | $51M + $30M debt | Feb 2023 | Princeville Capital, Citi Ventures |
Figures compiled from public reporting; some round details are approximate.
Paul Dix and Todd Persen start a monitoring startup in San Francisco.
The team goes through YC and pivots toward the open source time series engine it built internally.
Mayfield Fund and Trinity Ventures lead the first major raise.
The company renames itself and releases the Telegraf collector.
Kaplan takes the helm; InfluxData raises a $16M Series B.
The Rust-rebuilt suite is announced alongside an $81M Series E round.
The new engine reaches general availability for self-managed deployments.
InfluxDB 3.6 adds natural-language querying; InfluxData partners with Litmus for AI at the edge.
Creator of InfluxDB. Previously built software for Microsoft, Google, McAfee, and Thomson Reuters. Sets the technical direction.
Co-founded the company (as Errplane) with Dix in 2012.
Veteran enterprise-software executive who scaled InfluxData from open-source project to commercial platform.
The company is remote-first and open-source-native. Its stated values run to diversity and respect, action-oriented ownership, humility as a driver of learning, calculated risk-taking - treating failure as a tool for innovation - and active participation in open source communities.
Partners with Litmus for scalable industrial AI; Telegraf Enterprise and Telegraf Controller v1.0 reach GA.
Explorer gains a natural-language query beta and an expanded Python Processing Engine.
InfluxDB 3 Core and Enterprise become generally available for self-managed use.
The Apache Arrow-based, Rust-core product suite is announced.
InfluxData builds InfluxDB, an open source time series database and platform for collecting, storing, and analyzing timestamped data from sensors, applications, infrastructure, and IoT devices in real time.
It was founded in 2012 in San Francisco by Paul Dix and Todd Persen, originally as a monitoring startup called Errplane. Paul Dix created InfluxDB and remains CTO; Evan Kaplan has been CEO since 2016.
InfluxDB 3 is a complete rewrite in Rust using the FDAP stack - Flight, DataFusion, Apache Arrow, and Parquet - enabling unlimited cardinality, cheaper object storage, and SQL querying.
It uses an open-core model: the database and Telegraf collector are free and open source, while InfluxData sells managed cloud services, self-managed Enterprise and Clustered editions, and support subscriptions.
More than 1,900 paying customers and over a million active instances, including Tesla, IBM, Cisco, PayPal, Nordstrom, NVIDIA, and Siemens, across DevOps, industrial IoT, and financial services.