Published at Mar 6, 2024
InfluxDB Alternative
Written by Carlota Soto
Time-series database InfluxDB has gained substantial traction among developers and companies in the past few years. However, Influx Data, the company behind the database, has made questionable product decisions that have recently stirred some waves in developer communities.
In this article, we discuss some InfluxDB alternatives you might want to consider if you’re looking for another database for your time-series data. As the developers of Timescale, we believe that Timescale is a great alternative to InfluxDB—let us share eight reasons why we think you should consider giving it a try.
The first reason is straightforward. PostgreSQL is the most loved database by developers, with a user base that continues to grow and a vibrant community. With its 35+ years of development, using PostgreSQL is simply a smart strategic choice. Specialized databases tend to come and go, but PostgreSQL stays.
Let’s be honest: databases always lock you to some degree—nobody wants to migrate databases every year. If you’re forced to link the success of your application to a database, make it PostgreSQL, a technology that you know you can trust (and that’s not going anywhere).
Timescale is not compatible with PostgreSQL—it is PostgreSQL. Our core database, TimescaleDB, is built as a PostgreSQL extension, meaning that the experience of working with Timescale is exactly the same as working with PostgreSQL—with a performance boost for your time-series data. When you or your applications interact with a Timescale database, they interact with a PostgreSQL database.
Great practical advantages come from the fact that Timescale is PostgreSQL under the hood. For example, if you’re already using PostgreSQL, your “time-series database experience” will be limited to your time-series tables (called hypertables), which you can store in the same database as the rest of your data.
In your hypertables, you’ll get the extra performance you’re looking for for your time series (via automatic partitioning, optimized materialized views, query planner improvements, columnar compression, and many other things) without affecting the functionality of your regular PostgreSQL tables.
This interaction summarizes it well:
What Timescale allows you to work with a unified database environment where time series and regular data coexist and complement each other. Instead of resigning to the operational overhead needed to manage, secure, and maintain two distinct database systems, you can use Timescale and simply work two different table types—regular PostgreSQL tables and hypertables—both happily living in the same database.
To bring this home, let’s build on a scenario inspired by the user above. Consider an application with a PostgreSQL database containing a `user` table and a `device_reading` table that stores time-series data from sensor readings. With TimescaleDB, these tables can effortlessly coexist, and you can perform SQL queries that join data from both tables, thanks to the unified environment.
-- A typical users table
CREATE TABLE users (
user_id SERIAL PRIMARY KEY,
username TEXT NOT NULL,
email TEXT NOT NULL
);
-- A Timescale hypertable for device readings
CREATE TABLE device_reading (
time TIMESTAMPTZ NOT NULL,
user_id INT NOT NULL,
reading FLOAT NOT NULL
);
-- Convert the device_reading table into a hypertable
SELECT create_hypertable('device_reading', 'time');
-- Example of a query that joins data from the users and device_reading tables
SELECT users.username, avg(device_reading.reading) as average_reading
FROM users
INNER JOIN device_reading ON users.user_id = device_reading.user_id
WHERE device_reading.time > NOW() - INTERVAL '1 week'
GROUP BY users.username;
Another direct advantage of using PostgreSQL is its query language. SQL has a rich history, is well-documented, stable and the third most commonly used language among professional developers.
Have a question on how to structure a query? Millions of developers in SQL communities worldwide are ready to help you. Do you need to add a new tool to your stack? You can turn to the entire SQL ecosystem of third-party tools, connectors, and visualization options to easily set up this integration.
Learning a new, specialized query language (that may even be good!) to interact with one product (and having to adapt your code afterward) is not a very productive use of your time. Especially when that query language might change between versions of that particular product, like the particular case of InfluxDB.
In our experience (which, to be fair, mostly relates to the 1.x version of InfluxDB), Timescale outperforms InfluxDB in most queries:
In our testing, for simple rollups (i.e., GROUPBYs), Timescale exhibited 460 % of the performance of InfluxDB on configurations with 100 and 4,000 devices with 10 unique metrics generated every read interval.
Timescale also demonstrated 168 % of the performance of InfluxDB when aggregating eight metrics across 100 devices and 156 % when aggregating eight metrics across 4,000 devices.
For double rollups aggregating metrics by time and another dimension (e.g., GROUPBY time, deviceId), InfluxDB showed better performance than TimescaleDB when aggregating one metric. But, as the number of aggregated metrics increased, TimescaleDB achieved 188% better performance than InfluxDB.
For complex queries, TimescaleDB vastly outperforms InfluxDB and supports a broader range of query types; the difference here is often in the range of seconds to tens of seconds, with Timescale displaying a 344-7,100 % performance improvement over InfluxDB.
This last point is worth emphasizing. Having a fast database is key to building great applications for your users, but this also demands query flexibility from your database. SQL offers rich query functionality that is both familiar and powerful—the full SQL features (combined with Timescale’s library of SQL functions) allow you to write time-based analytical queries in a few lines of code and run them with excellent performance, something that traditional time-series databases like InfluxDB will struggle to do.
The essence of a reliable database lies in its promise to safeguard data integrity. While InfluxDB grapples with the herculean task of architecting reliability from scratch, Timescale can leverage the decades of hard, careful engineering work that the entire PostgreSQL community has done to build a rock-solid database that supports millions of mission-critical applications worldwide.
Getting all the corner cases right when building a database is extremely hard: every database goes through a period when things get perfected from real-world experience. The significant advantage of PostgreSQL is that it went through this period in the 1990s, while InfluxDB is still figuring things out today.
Given Timescale’s design, we’re able to leverage the full spectrum of tools that the PostgreSQL ecosystem offers and has rigorously tested, including streaming replication for high availability and read replicas, pg_basebackup, and log shipping/streaming for incremental backups and arbitrary point-in-time recovery, pgBackrest or WAL-E for continuous archiving to cloud storage, pgBouncer for connection pooling, and many more.
One of the essential features that Timescale enables in your hypertables is columnar compression. Compression significantly reduces the storage footprint of time-series data, ensuring you can manage large data volumes without seeing a proportionate rise in storage costs. This is important because time-series data accumulates very quickly—you may not be paying much for storage today, but this will change as soon as your application grows.
Our approach to adding columnar compression to PostgreSQL is both innovative and practical, transforming the traditional row-based storage of Postgres tables into a columnar format. Using specialized compression algorithms depending on each data type, Timescale achieves compression rates of over 90 %. The best part? Queries also get faster over compressed data, especially analytical queries over large volumes that benefit from a columnar data structure.
"Compression was a game-changer from our perspective: not having to worry about getting databases on the order of 5, 10, or 15 TB to store this information was a massive factor for us [… ]. For one of our larger customers, we normally store about 64 GB of uncompressed data per day. With compression, we’ve seen, on average, a 97 percent reduction.” (Source)
Carrying on the topic of pricing experience, you don’t want your database bill to grow out of control, and you don’t want it to be surprising, either. At Timescale, we have a basic pricing structure without pricing gotchas or hidden costs: at the end of the month, you’re charged for the storage you’ve used and the compute you’ve provisioned. That’s it.
Our experience with InfluxDB Cloud was quite different. We set up a small service to do some benchmarking, and a bill that we expected to be around $10-20 was closer to $2,000. We still don’t know exactly why.
Lastly, our world-class Support and Customer Success teams might be the best reason to try Timescale. We carefully monitor customer satisfaction, and our scores are very close to 100 % customer satisfaction every time. These are only a few examples of the praises we hear from our customers regularly:
"I have nothing but amazing things to say about your Customer Support services.”
"Support has been bar none. We look forward to migrating our infrastructure to Timescale and are excited about the support we will receive in the process.”
"The interaction with Support went far beyond my expectations.”
"I am very happy with the experience I’ve had and the support Timescale has been able to provide. This is a symbiotic relationship: you provide solutions to me, I provide insights to make your product better.”
“You could have sufficed by saying, “read the PostgreSQL manual”—but I appreciated that you gave me a few commands.”
"Timescale support is A+.”
As a Timescale customer, you can sit down with an expert engineer who cares about making you successful, not just fixing TimescaleDB issues. They’ll advise you on database design, query optimization, and everything in between. When you look good, we look good.
The decision to move from InfluxDB to Timescale is yours entirely. To make your decision, it’ll help you to evaluate your priorities, together with the requirements of your use case today and in the future. Picking a database is like picking a partner: you’ll be together for years, so take your time!
If we piqued your curiosity, give Timescale a try. If you're already running a PostgreSQL database on your own hardware, you can simply add the TimescaleDB extension. If you prefer to try Timescale in AWS, create a free account on our platform. It only takes a couple of seconds, no credit card required.