A look back at 2018: Here’s what we accomplished in Timescale’s 2nd year

Over 1 million downloads, 15 releases including 1.0, production deployments worldwide, the launch of the TSBS, and the release of the Timescale License

Less than two years have passed since we first launched TimescaleDB to the world. That first year was an exciting ride: we connected with thousands of developers at events worldwide, experienced enthusiastic enterprise adoption and production deployments, and published several major releases.

2018 was an even bigger year for the team at Timescale, with more concrete milestones and major developments, including TimescaleDB version 1.0! While there’s too much to include in just one blog post, we’ll highlight some of those milestones and provide a taste of what’s to come in 2019.

To TimescaleDB users worldwide, from our earliest supporters to the folks who joined our Slack community just this week, thank you! Your feedback and support motivates us every day.

And for those of you still considering TimescaleDB, you can install it today via Docker, various Linux Package managers (Ubuntu, Debian, RedHat, etc.), Homebrew (MacOS), a Windows Installer, and an Amazon AMI. (And you can get started with these tutorials.)

Major Milestones

Beginning in January 2018, we announced our $16M Series A financing from Benchmark, NEA, and Two Sigma Ventures. With this financial support, we have been able to triple the size of our team, hiring a distributed team of hardworking, talented business minds and engineers in New York and across the globe in places like Sweden, Macedonia, Moscow, Germany, Portland, SF, and LA. Together we’re spreading the word about and accelerating the development of TimescaleDB. If building the best all-purpose time-series database system sounds exciting to you, please get in touch — we’re hiring!

More recently, back in September, we announced version 1.0. The release of version 1.0 signified that TimescaleDB is officially not just production-ready, but also enterprise-ready. (TimescaleDB was production-ready very early on, owing to the PostgreSQL foundation it is built on. In fact, 4 months after our launch, TimescaleDB was deployed in production to power the operational dashboards monitoring dozens of hydroelectric power plants around the world.)

And just two weeks ago we announced the Timescale License (TSL) for some upcoming new features. (Note: our existing code base, as well as the core of our database going forward, remains open-source under Apache 2.) We believe the TSL is a necessary step to ensure the long-term sustainability of Timescale as a business and, subsequently, the development of the open-source core of TimescaleDB.

We also traveled throughout the world, delivering numerous technical talks and our first public training. Our R&D team published numerous, thorough technical benchmarks against other systems (see a list of them below) and open-sourced the Time Series Benchmark Suite, a collection of Go programs that are used to streamline the benchmarking of time-series databases and intended to standardize the process.

And so, to enumerate the other moments, highlights, whereabouts, and welcomings in 2018, we’ve listed them below. (Skip the bottom for what’s to come in the New Year.)

Top 5 Blog Posts of 2018

In 2018, we published 28 long-form blog posts that included technical benchmarks, write-ups on product improvements and optimizations, and thought pieces about the OSS industry.

Here are the Top 5 posts from the Timescale Blog in 2018:

1. TimescaleDB vs. InfluxDB: purpose-built differently for time-series data
2. High Availability and scalable reads in PostgreSQL (psst! HA comes free in TimescaleDB)
3. Announcing TimescaleDB 1.0: First enterprise-ready TSDB to support full SQL & scale
4. Eye or the Tiger: Benchmarking Cassandra vs. TimescaleDB for time-series data
5. Why SQL is beating NoSQL, and what this means for the future of data

Top Write Ups from the Timescale Community

• Why DNSFilter replaced InfluxDB with TimescaleDB — Mike Schroll, CTO
• Time-Series at ShiftLeft — Preetam Jinka, Senior Software Engineer
• Using TimescaleDB with Hasura GraphQL — Anon Ray, Engineer
• Store application logs in TimescaleDB — Komu Wairagu, Developer at Jumo

Community Growth & Adoption

Support!

• 1,800+ members in the TimescaleDB Slack Community & Support Group
• We’ve held over 20 Open Office Hours sessions to date

Launched the Timescale webinar series

• RSVP for: Getting Started with TimescaleDB and Defining Best Practices (scheduled for Jan 29)
• Watch: Performant time-series data management with PostgreSQL
• Watch: Visualizing Time-Series Data with Grafana & TimescaleDB

Events & Presentations

• Exhibited at 20+ conferences throughout the US, Europe, and India
• The IoT NY Meetup has grown to over 6,300 members
• Launched the Time-Series Data NY Meetup
• Watch presentations from the events above and more on our YouTube

Sticking close to our roots (PostgreSQL & Open Source)

We wrote about PostgreSQL and how awesome it is, our thoughts on open-source businesses and the state of the OSS industry.

PostgreSQL:

• Why the European Space Agency uses PostgreSQL
• Evaluating High Availability Solutions for TimescaleDB + PostgreSQL
• Postgres as the Sub-Structure for IoT and the Next Wave of Computing

Open source:

• Open-source and the demise of proprietary software
• 5 ways open source software companies make money

Performance Benchmarks

As part of our engineering efforts (detailed more below) we published thorough technical benchmarks and comparisons of TimescaleDB vs. alternative database solutions. While running these benchmarks it became clear that there lacked a standard suite of tools for benchmarking time-series data storage solutions and so we released the Time Series Benchmark Suite.

How TimescaleDB compares against…

• InfluxDB (plus an analysis on SQL vs. Flux)
• Cassandra
• MongoDB
• Introducing the Time Series Benchmark Suite (TSBS)

Ecosystem Partnerships & Integrations

Since Timescale’s inception, we’ve always stressed the power of ecosystem compatibility. As a Postgres extension, TimescaleDB integrates with the entire Postgres ecosystem of tools and connectors. That said, there have been cases where the behavior of an integration is sub-optimal. When we identify these gaps, or more often when our users identify these gaps for us, we work hard to ensure a more positive user experience and fix the problem. Some of those improvements and first-class integrations are highlighted here:

• Prometheus: Timescale engineers develop the pg-prometheus PostgreSQL extension
• Grafana: Timescale engineers develop the PostgreSQL/TimescaleDB datasource
• Seeq: Use TimescaleDB as a backend for your industrial monitoring
• Hasura: GraphQL queries for PostgreSQL (and TimescaleDB)
• Sensu: Send metrics directly to TimescaleDB (developed by the Sensu team)
• Mode: SQL reporting, with support for TimescaleDB as a backend
• DBeaver: an alternative to pgadmin
• (And of course, everything else in the PostgreSQL ecosystem)

In 2019 we’re excited to expand our ecosystem via even more partnerships and integrations!

Over 15 releases

In 2018, our engineering team really accelerated the development of TimescaleDB. The two main focuses here were refining the overall user experience, from installation through production deployment, and making stability improvements on the way to version 1.0. Summaries for many of these releases are below.

For more details on each of the releases, please click the links:

• 1.1.0 (2018–12–13): Beta support for Postgres 11. Several performance optimizations aimed at improving chunk exclusion for read queries. Packaging of timescaledb-tune tool (currently in beta) with our Debian and Linux releases.
• 1.0.1 (2018–12–5): Optimizations for 1.0
• 1.0.0 (2018–10–30): Broad usability improvements, added foundation for background job automation and scheduling, increased database hardening and testing, native support for Grafana and Prometheus. (RCs: 1.0.0-rc3 (10–18), 1.0.0-rc2 (09–28), 1.0.0-rc1 (09–12))
• 0.12.1 (2018–09–19): Improved testing, code cleanup, and fixed other issues.
• 0.12.0 (2018–09–10): Introduced a background job framework and scheduler.
• 0.11.0 (2018–08–08): Continued hardening to address a number of less frequently used schema modifications, functions, or constraints.
• 0.10.1 (2018–07–12): Improved memory management, fixed handling of dropping triggers, corrected handling of table and schema name changes to chunks.
• 0.10.0 (2018–06–27): Planning time improvement, smarter use of HashAggregate, new convenience function.
• 0.9.2 (2018–05–04): Fixed handling of ‘DISCARD ALL’ command when parallel workers are involved, fixed user permission bug.
• 0.9.1 (2018–03–26): Updated preloader, added ABI compatibility.
• 0.9.0 (2018–03–05): Support for multiple extensions, streamlined update process, significant refactoring to simplify and improve codebase.

Looking ahead to 2019: A Year of Moonshots

For Timescale, 2018 was a year of milestones but as we look ahead to 2019, we’re preparing for a year of moonshots. We’re planning even more events to meet with you face to face. We’ve ironed out our product roadmap, and are actively developing scale-out clustering, adding a number of enterprise-grade features, and bringing TimescaleDB to a cloud near you.

In 2019 we will continue our constant pursuit of this mission: to help organizations harness the power of time. By helping organizations capture, store, and analyze data over time, also known as time-series data, we enable them to understand the past, monitor the present, and predict the future. We believe that time-series data analysis is foundational to a diverse set of applications and use cases, and that, fundamentally, all data is time-series data. As a result, we believe that we are building a foundational technology for the future of data and computing. We hope you join us for the ride.