Canner vs. Google Cloud SQL for PostgreSQL
Cloud SQL is Google’s managed Postgres. Canner is the Canadian-owned one that sits next to your app.
Cloud SQL brings high-availability instances, logical replication and the Google Cloud ecosystem. Canner gives you Postgres on every plan, in one Canadian-owned platform with your app, cron and storage, priced flat in CAD. Neither is right for every workload; here is where each one wins.
- Cloud SQL is stronger on high availability, scale and integration with the rest of Google Cloud. Canner has one region, no multi-zone failover and no read replicas.
- Where Canner differs: Canadian ownership. The Government of Canada’s own white paper says data in the cloud may be subject to the laws of other countries, and that a US-law organization can be compelled regardless of the data’s location. Google is US-incorporated; Canner is not.
- On Canner the database, the app, cron jobs and object storage live in one place with flat CAD pricing. On Google Cloud the database is one metered service among many, and network egress is billed on top.
| Feature | Canner | Google Cloud SQL for PostgreSQL |
|---|---|---|
| Data location | Montreal, Quebec — the only region | Many Google Cloud regions; check Google’s documentation for the ones that offer Cloud SQL |
| Jurisdiction | Canadian-owned, no US affiliate; not subject to the US CLOUD Act | Google is a US-incorporated company. The Government of Canada’s own guidance says location does not remove exposure to foreign law: a US-law organization can be compelled “regardless of the data’s location” † |
| Managed Postgres | On every plan, including the free Starter: an isolated database and role per project (one on Starter, several on paid plans), DATABASE_URL injected. Postgres 17 or 18, chosen at creation | Separate metered service (instance, storage and network billed individually); see Google’s pricing page for current rates † |
| Pricing model | Flat CAD: Starter free, Live CA$9/month, Studio CA$149/month. Storage is one pooled limit per organization (5 / 25 / 100 GB); +50 GB blocks at CA$5/month | Metered by resource. Internet egress is listed at $0.19/GiB † |
| High availability | None. One region, one provider; no multi-zone failover, no automatic failover and no read replicas | High-availability instances exist (they cannot be logical-replication subscribers) † |
| Backups & point-in-time recovery | Scheduled encrypted backups with the retention you set, manual backups, verification with email alerts, restore into a new database or in place, and copies to up to 5 of your own S3-compatible buckets. Point-in-time recovery (1–35 days, into a new database) with the Dedicated instance add-on, CA$9/month per project | Point-in-time recovery is supported; a recovery always creates a new instance † |
| Extensions | With the Dedicated instance add-on: pgvector, PostGIS, pg_trgm, pg_cron, hstore and others (14 in the catalog). No fdw or dblink | pgvector, PostGIS and pg_cron supported † |
| Logical replication | With the Dedicated instance add-on: replication out to your own subscribers, and following an external server live for a no-downtime migration (never as a superuser). Physical replication is closed | Publisher and subscriber supported; high-availability instances can’t be subscribers † |
| Moving in and out | Migrate-in streams from any public Postgres server; pg_dump export, migrate-out and copy on every plan. `canner db migrate-in` from the CLI | See vendor documentation |
| What sits beside the database | Your app, cron jobs, object storage, scheduled SQL, cookieless analytics and a CLI, REST API and MCP server, in one dashboard | The database only. Compute, networking, identity, storage and scheduling are separate Google Cloud services you assemble |
| Ecosystem & enterprise fit | A focused platform; Enterprise from CA$600/month by quote | Deep integration with Google Cloud’s IAM, networking and data services, and a very large ecosystem |
† Competitor details come from each vendor's own published pages on the dates shown; prices and plans change, so check the vendor before you decide. Jurisdiction (2026-09-26) · Managed Postgres (2026-09-26) · Pricing model (2026-09-26) · High availability (2026-09-26) · Backups & point-in-time recovery (2026-09-26) · Extensions (2026-09-26) · Logical replication (2026-09-26)
When Canner is the right call
You are shipping a Canadian app or internal tool and want the database beside it: one place, one bill in CAD, Postgres on every plan, and no project, VPC or instance sizing to work out first. It matters most when your buyer asks about jurisdiction. Google is US-incorporated, and the Government of Canada’s own guidance says the location of the data does not remove that exposure. Canner is Canadian-owned, hosted only in Montreal, and not subject to the US CLOUD Act. You also get backups you can verify and copy to your own buckets, and point-in-time recovery on a dedicated instance for CA$9/month.
When Google Cloud SQL still makes sense
If you need high-availability instances, a database that grows well beyond one server, or tight integration with the rest of Google Cloud (its IAM, networking, analytics and AI services), Cloud SQL is the stronger product and Canner does not try to match it: Canner has a single region, no automatic failover and no read replicas. It is also the natural pick when your estate is already on Google Cloud, or when an enterprise agreement and a platform team exist. If Google Cloud satisfies your residency requirement and nothing more, Cloud SQL is a sound choice.
Ready to give it a try?
Free tier available — Starter is permanent, no card required to sign up. Cancel anytime.