QuestDB Cloud Is Gone. Here's Where to Run QuestDB Now.
If you are landing here from a dead QuestDB Cloud bookmark, the short version: the self-serve managed product is discontinued, the AWS Marketplace listing is inactive, and the official path forward is QuestDB Enterprise/BYOC, which sits behind a sales conversation. As of August 2026 there is exactly one self-serve managed QuestDB on the market, and it is Layerbase. This post lays out the options honestly, including the ones that are not us.
To be clear up front: nothing is wrong with QuestDB the database. The column store is excellent, ingestion over InfluxDB Line Protocol is genuinely fast, and the Postgres wire compatibility means your existing SQL tooling mostly just works. What disappeared is the easy on-ramp, not the engine.
Your three options in 2026
Self-host. Provision a VM, install QuestDB from the tarball or Docker image, and own TLS, backups, monitoring, JVM tuning, and file-descriptor limits yourself. Cheap at low traffic, and completely reasonable if you have the ops appetite. The QuestDB deployment docs are good.
QuestDB Enterprise/BYOC. You bring an AWS or Azure account (GCP is still listed as coming soon), QuestDB manages a deployment inside it, and you get replication, RBAC, and support SLAs under a contract. If you are a company with a procurement process and a production time-series workload measured in serious money, this is probably your answer, and you should email them. Pricing is not public, so we will not guess at it.
The Enterprise line is also where the engineering is going, which is worth saying plainly because it is a real argument for the contract. Enterprise 3.3.1 landed on June 9, 2026 with storage policies, custom CA support, and finer-grained access control, all of it aimed at teams with compliance reviewers. Enterprise 3.3.2 followed on June 23 with Parquet promoted to a full table format rather than an export target, native ARM64 Linux binaries, and a single-version QWP wire protocol that removes a chunk of client-compatibility guesswork. None of that is bad news. It is just gated: there is still no signup form, no public price, and no way to try it without booking a demo first. If your evaluation has to start this afternoon, the gate is the problem, not the roadmap.
Layerbase Cloud. Self-serve managed QuestDB: sign up, click create, connection details in a couple of minutes. The rest of this post covers what that actually gets you.
What QuestDB on Layerbase looks like
Both ingestion paths, exposed from the dashboard. QuestDB accepts InfluxDB Line Protocol for high-throughput writes and speaks the Postgres wire protocol for SQL queries. Your database's dashboard page shows both endpoints, so Telegraf-style ILP pipelines and psql-style query tooling point at the same instance the way they did on QuestDB Cloud.
A standard Postgres-wire endpoint for queries. psql, the pg npm package, psycopg, pgx, Grafana's Postgres data source: anything that speaks Postgres wire runs your SAMPLE BY and LATEST ON queries. No custom driver.
A web SQL console. The dashboard includes a query console, so checking an ingest rate or eyeballing a table does not require a local client at all.
Always-on, not scale-to-zero. QuestDB is one of our Performance-class engines. It runs on reserved capacity from your plan's pool rather than sleeping when idle, so there is no JVM cold start in front of your first morning query. First-time provisioning takes about 1 to 2 minutes (the binary is large); after that it stays up.
The honest pricing note
QuestDB is a Performance-class engine, which means it is not on the Layerbase free tier. It requires the Pro plan: $15/month flat, up to 10 databases, all engines, with a 1.5 GB reserved RAM pool you allocate however you like and $10/month blocks (+1 GB RAM, +1 vCPU, +25 GB storage) if you need more. There are no usage meters on any plan, so a heavy ingest month costs the same as a quiet one. Pro comes with a 7-day free trial, which is enough time to run a real ingest test before any money moves.
Is $15/month more than free? Yes. Is it less than an always-on VM you also have to operate, or an enterprise contract you have to negotiate before writing a row? Also yes. For the evaluation-to-small-production range that QuestDB Cloud used to serve, that is the slot we are filling.
Deciding quickly
- Production workload, real budget, want SLAs and replication: talk to QuestDB about Enterprise/BYOC.
- Comfortable operating your own VM and want the lowest cash cost: self-host.
- Want managed QuestDB today without a sales call: create one on Layerbase and be writing rows inside ten minutes.
We wrote a broader survey of the post-shutdown landscape, including local options for development, in QuestDB Cloud alternatives, and there is a hands-on ingest walkthrough in Getting started with QuestDB. The engine details, FAQ, and connection specifics live on the QuestDB engine page.
QuestDB deserved a better fate than sales-gated-only. Until that changes upstream, this is where you run it.
Keep reading
- QuestDB 9.4: Time-Series Partitions You Can Store as ParquetQuestDB 9.4 lets a table declare Parquet as its partition format at CREATE TABLE time. Here is what Parquet actually is, why a time-partitioned table is the ideal shape for it, and how the feature behaves on a real database.
- QuestDB Cloud is gone. Here is what to use instead.QuestDB retired its self-serve cloud for Enterprise BYOC. If you want a managed QuestDB instance without a sales call, here are the options that still exist.
- Getting Started with QuestDBBuild a time-series sensor pipeline with QuestDB and TypeScript, learn why SAMPLE BY beats verbose GROUP BY queries, and run it all in one script.
- Best Database for IoTA practical guide to picking the right database for IoT sensor data, covering QuestDB, InfluxDB, and PostgreSQL with SpinDB examples you can run locally.