Tiger Cloud

TimescaleDB is an open-source project. You can choose to self-host it or use managed hosting at Tiger Cloud by the creators of TimescaleDB with $1,000 free credits. But wouldn't it make more sense to let experts manage everything so you can focus on your main goal? You want to analyze data, not manage servers and PostgreSQL! And when using Tiger Cloud, you also get many features on top of TimescaleDB that require complex infrastructure beyond PostgreSQL.

Do you really want to manage all that?

Configuring a production-grade PostgreSQL setup can take many days (or even weeks). It is also ongoing work, not just a one-time task! Do you want to take full responsibility for it? A database outage can vary from a minor inconvenience to completely halting work for everyone. Would you prefer to be the one blamed for a problem or to have the option to pass the blame to someone else?

The list of things you need to set up, manage, and monitor is long. The most important tasks you need to do (but there are more):

• Server Management: You need to manage a Linux server and secure it against attackers. Any announced security vulnerabilities must be patched immediately. Additionally, you must install, configure, update and optimize PostgreSQL to align with your server's capabilities.
• Performance: You need monitoring to be notified when processes are using too much CPU, when the filesystem or disks have issues that cause slow performance, or when any other part of Linux or your hardware is not working properly. Obviously, you also need to know how to handle these situations. On top of that, you should keep an eye on your PostgreSQL database and all its health metrics. This includes dealing with table bloat, lock waits, and slow queries. A query that has been fast for months can become slow, or a new query might overwhelm the server and slow down every other query.
• Backups: You don't want to lose any data. Never! You must implement a backup procedure and regularly test it. Many companies discovered that their backups failed or were corrupted for months only when they needed them. This is where mistakes can have serious consequences: While a database being unavailable for a few hours can be tolerated, losing all data can end a business.
• Availability: Your database should never be unreachable for more than a few seconds: Installing patches, rebooting a machine, failed configuration updates, a datacenter being unreachable and a (virtual) server or disk failure shouldn't be a major concern. You must build a PostgreSQL cluster across multiple servers in different availability zones with automatic failover. Unfortunately, this setup is still quite complex today.

When you use Tiger Cloud, everything is taken care of for you. The servers and PostgreSQL are automatically managed, tuned and monitored. With the Performance Insights feature, you can see which queries are currently blocked because they're waiting on a lock held by another query. You can also view all slow queries along with their past execution details such as time, rows loaded, and cache hit rate. Incremental backups are saved continuously, so you can restore your database to the minute before you e.g. accidentally deleted data. Additionally, with a single click, you can upgrade your PostgreSQL database to a multi-AZ cluster or add read replicas for improved performance.

It is simple to migrate from your self-managed setup: You can set up Tiger Cloud's LiveSync feature to sync your database with TigerCloud in just a few minutes. All data migration occurs in the background, and you can decide when to switch over. There's no need to rush. You can let it sync for days or even weeks until you're ready.

Faster than your Database

All TimescaleDB performance optimizations are available to you with the self-hosted version. Nothing is locked behind a paid tier. So Tiger Cloud isn't faster at running queries; it's faster at stopping you from being slow:

• You get automatic tuning of your PostgreSQL settings based on the resources available to your Tiger Cloud instance. The chunking intervals for your hypertables are also automatically adjusted to keep every chunk at a reasonable size. So when your product grows or usage changes, you won't experience a steady performance degradation that often goes unnoticed until performance drops significantly.
• The more complex a query is, the longer it takes to execute and the more server resources are used. This impacts your database's performance and affects other read and write queries. As a result, your pipeline for saving new rows slows down, the queue grows and your queries produce outdated results. On Tiger Cloud, you can create read replicas with just one click. You then run heavy queries there without impacting anything else.
• The load on a database changes over time. When the database grows, you may need a bigger machine. However, shutting down a database, picking a new instance type and restarting it takes everything offline for a few minutes. That's not perfect. With Tiger Cloud's dynamic compute resizing, you no longer have to worry: You choose a new instance size and all high-availability replicas are upgraded first. Then, your primary database switches to an upgraded high-availability replica. That's it. There's no downtime at all. Your database clients lose their connection but will reconnect instantly. With such an easy process, you can change the instance size whenever you need more resources for large backfills or a big spike in usage.

More Features

All features of TimescaleDB are open-source. That's the core belief of Tiger Data when creating the PostgreSQL extension. However, some features go beyond what PostgreSQL can offer and need many infrastructure integrations to work. These are only available in the cloud, e.g. Tiered Storage. There's more, like Terraform support to set up your database with code and Insights to identify slow queries. But the most exciting parts are the data connectors and support for agentic workflows.

Data Connectors

At some point, writing to a TimescaleDB hypertable will become a bottleneck for you: Inserting single rows directly when you have new data limits your maximum rows per second drastically because it is slow, whether on normal tables or hypertables. You need to implement complex batching by writing data somewhere else and then inserting rows in larger batches: This is a lot of work. Tiger Cloud has solved this for you:

• You can send all your new data to Kafka as soon as you receive it. It's optimized for this purpose. With the deep integration for Confluent Cloud Kafka clusters, you only need to configure a mapping of your columns in the Kafka schema registry. Tiger Cloud will pull all data from the Kafka stream, batch it and insert it into your TimescaleDB database. You don't have to manage Kafka offsets, implement retry logic or write anything to distribute work across many workers.
• If you already have the data in a table or stored elsewhere, you can use S3 syncing. For example, export all rows inserted in the past five minutes to a new S3 file every five minutes. Tiger Cloud will find it and start syncing automatically. It's that simple. You can leave the current logic for storing new data untouched and only add an unrelated periodic export job to fill your TimescaleDB hypertables.

But TigerCloud also helps when you need to analyze your data with an analytics platform that can't (yet) use TimescaleDB. All your changes to a hypertable can be automatically synced to Apache Iceberg in Amazon S3 Tables or via the Iceberg REST catalog. Any software that can read Iceberg data now works with your TimescaleDB data.

Agentic Workflows

AI has simplified many development tasks that were very time-consuming before. Considering a migration of an existing project to TimescaleDB for better performance? Testing this used to take days. Tuning an existing setup with a different schema, ingestion pipeline or queries? AI can suggest many recommendations. But which of these will actually make a difference?

Tiger Cloud has an MCP server that lets your AI agent quickly spin up new databases for experimentation. Testing the TimescaleDB performance then means just having an agent work overnight to import your data and update your application. Once you've already moved all your data to Tiger Cloud, an agent can create forks of your live database with all the data in just seconds. Thanks to zero-copy forks using Fluid Storage, this is always fast no matter how big the database is. So you can spawn a fleet of agents that run independently on their own database forks to test performance optimizations and implement them only if they perform well.

Let's jump over to Tiger Cloud and create your account with $1,000 free credits. They will manage your database and you can create a production-ready setup in minutes. Point your AI agent at the course, let it migrate all data overnight and you have a fast database tomorrow when you get your morning coffee.