PITR
Point-in-time recovery to any minute in the window
DatabasesFrom $29/app/month
Managed PostgreSQL Hosting
Advanced open-source relational database
What is PostgreSQL on ManageStacks?
PostgreSQL on ManageStacks is production-grade Postgres deployed to your own AWS, Azure, or GCP region — priced flat at $29 per database per month regardless of storage size, with WAL archiving + point-in-time recovery, PgBouncer connection pooling, streaming replication for HA, and automated major-version upgrades tested first on a staging clone. Materially cheaper than RDS or Cloud SQL once you're past a few GB, and your database stays in your cloud region.
PostgreSQL on ManageStacks is production-grade Postgres deployed to your own AWS, Azure, or GCP region — priced flat at $29 per database per month regardless of storage size, with WAL archiving + point-in-time recovery, PgBouncer connection pooling, streaming replication for HA, and automated major-version upgrades tested first on a staging clone. Materially cheaper than RDS or Cloud SQL once you're past a few GB, and your database stays in your cloud region.
Last updated Official siteSource on GitHubDocumentation
About PostgreSQL
What PostgreSQL does, and why teams deploy it.
PostgreSQL is the world's most advanced open-source relational database. Known for its reliability, data integrity, and SQL-standard compliance, Postgres handles workloads from single-tenant SaaS backends to multi-TB analytical warehouses with hundreds of concurrent connections.
Postgres supports advanced data types (JSON/JSONB, arrays, hstore, ranges), sophisticated indexing (B-tree, GIN, GiST, BRIN, hash), full-text search, geospatial via PostGIS, logical + streaming replication, and native partitioning for large tables. Its extension system (pg_stat_statements, pg_cron, pgvector, TimescaleDB, Citus) is unmatched in the open-source database world.
Running production Postgres yourself means running the postgres process, configuring shared_buffers/work_mem/max_connections correctly for your workload, setting up WAL archiving for PITR, running PgBouncer in front (Postgres processes are heavy — you don't want to run out of them under connection storms), scripting nightly vacuum + reindex, and testing pg_upgrade before every major-version bump. ManageStacks handles all of that.
DIY vs ManageStacks
What running PostgreSQL yourself looks like — and what it looks like with us.
DIY self-hosting
- Install Postgres 16, tune postgresql.conf (shared_buffers, work_mem, max_connections) for your workload
- Set up WAL archiving to S3, verify PITR restores work, and script nightly vacuum + reindex
- Install and configure PgBouncer in front for connection pooling
- Test pg_upgrade for major versions in staging; validate no application-side breakage
- Build streaming replication + failover manually with Patroni or repmgr
On ManageStacks
- Subscribe through your AWS, Azure, or GCP marketplace
- Postgres comes up tuned for your instance size, with PgBouncer in front and WAL archiving on
- Prometheus-exporter metrics + Grafana dashboards ship pre-configured for connections, cache hit ratio, replication lag
- New Postgres versions get pg_upgrade-tested against a clone of your data before rollout
- Streaming replication + automatic failover included on HA plans
PostgreSQL on ManageStacks — key numbers
$29/mo
Flat per database, standard tier
HA
Streaming replication + automatic failover on Business+
pgvector
Vector search built-in, no separate vector DB
Key features
Everything PostgreSQL ships with, running on our stack.
- ACID-compliant with MVCC concurrency control
- Advanced indexing: B-tree, GIN, GiST, BRIN, hash, partial, expression
- JSON/JSONB with full indexing for semi-structured queries
- Logical and streaming replication with automatic failover on HA plans
- Native table partitioning for TB-scale tables
- Extensibility: pg_stat_statements, pg_cron, pgvector, PostGIS, TimescaleDB, Citus
- PgBouncer connection pooling in front (transaction and session modes)
- WAL archiving with point-in-time recovery to any minute
- Prometheus-exporter metrics + Grafana dashboards included
- Full data export — pg_dump, pg_dumpall, WAL archives on demand
How it deploys
From subscribe to live in minutes.
Subscribe
Subscribe to ManageStacks through your AWS, Azure, or GCP marketplace.
Provision
Postgres spins up with your chosen version, sized to instance, with PgBouncer, WAL archiving, and Grafana monitoring — typically 3-5 minutes.
Connect
Point your application at the PgBouncer endpoint (host/port/credentials shown in the dashboard). SSL is on by default.
Backup + monitor
Daily backups + PITR run in the background. Set up alerting on replication lag, connection count, or query latency through the pre-built Grafana dashboards.
Who this is for
Built for teams that want PostgreSQL to just work.
SaaS backends
You need a solid transactional database with PITR, HA, and clean upgrade discipline. RDS works but adds up. ManageStacks Postgres is flat-priced with the same operational shell.
Analytics + reporting workloads
You need Postgres for BI queries, dbt models, or a data mart. Add pg_partman for partitioning and pg_cron for scheduled aggregations. TimescaleDB and Citus available for time-series and horizontal scale.
AI / vector-search workloads
pgvector + pgvectorscale handle up to millions of embeddings with HNSW indexing — no separate vector database. Cheaper and simpler than Pinecone or Weaviate for most application-embedded semantic search.
Compliance & compatibility
What we handle, what PostgreSQL runs on.
Compliance & operations
- TLS-encrypted connections with certificate rotation
- WAL archiving to encrypted object storage in a separate region
- Point-in-time recovery within the retention window (30+ days)
- OS-level and Postgres security patches applied during your maintenance window
- GDPR data-residency — database stays in your chosen cloud region
- Row-level security policies (Postgres RLS) supported and used by application
Compatibility
- Version
- Postgres 14, 15, 16, or 17 (pin the major version you want)
- Runtime
- Postgres binary on containerised infrastructure
- Dependencies
- PgBouncer for connection pooling, S3/GCS/Blob for WAL archives
- Min. resources
- 1 vCPU / 2 GB RAM (dedicated, not shared)
How ManageStacks helps
We handle the parts you shouldn't be writing yourself.
ManageStacks deploys PostgreSQL with WAL archiving + point-in-time recovery, PgBouncer connection pooling, streaming replication, Prometheus-exporter metrics, and Grafana dashboards. We handle major-version upgrades on staging clones first, vacuum + reindex scheduling, and HA failover so your database is safe, tuned, and always available.
How it compares
PostgreSQL on ManageStacks vs the alternatives.
How Postgres on ManageStacks compares to the two dominant hyperscaler-managed Postgres options and running Postgres on your own VM.
| Property | Postgres on ManageStacksUs | AWS RDS Postgres | Google Cloud SQL Postgres | Self-hosted on your VM |
|---|---|---|---|---|
| Deployment | Managed on your AWS, Azure, or GCP | AWS-managed | GCP-managed | You provision + operate |
| Data residency | Your cloud region | AWS region | GCP region | Your cloud region |
| Pricing basis | Flat per database | Per instance-hour + storage + IOPS | Per instance-hour + storage | Your compute cost |
| PITR | Included (30+ days) | Included | Included | You configure WAL + PITR |
| HA / replication | Streaming replication + failover on HA | Multi-AZ / read replicas | HA config + read replicas | You configure Patroni/repmgr |
| Vector search | pgvector pre-installed | pgvector supported | pgvector supported | You install pgvector |
Comparison focuses on architectural properties (deployment model, pricing basis, open-source status) that don't change with vendor pricing pages. Verify current pricing on each vendor's own site.
FAQ
Common questions about PostgreSQL on ManageStacks.
Does ManageStacks support PostgreSQL point-in-time recovery?
Yes. WAL archiving is enabled by default with continuous archiving to object storage. Point-in-time recovery to any minute within the retention window (default 30 days, longer on Business/Enterprise) is available through the dashboard.
How does this compare to AWS RDS or Google Cloud SQL?
RDS and Cloud SQL are priced per-instance-hour + per-GB-storage + per-IOPS, which compounds fast as your database grows. ManageStacks is a flat $29 per database per month for the standard tier. For small-to-medium databases (< 500 GB), self-hosted on ManageStacks is materially cheaper. RDS is worth it for very large databases, complex HA topologies, or if you need vendor-specific features (Aurora, IAM auth). We handle the same platform operations — provisioning, backups, PITR, version upgrades, HA — but as a flat price and in your cloud account.
Can I set up PostgreSQL replication on ManageStacks?
Yes. Streaming replication with hot-standby replicas is available on all plans. Logical replication (for zero-downtime major-version upgrades or partial-table replication to a downstream system) is available on Business+. Automatic failover with a virtual IP is included on HA plans.
Does ManageStacks handle PostgreSQL major-version upgrades?
Yes. Major-version upgrades (14 → 15 → 16 → 17) use pg_upgrade with a staging clone first. We validate on your actual schema + data before rolling forward. Minor-version patches (16.2 → 16.3) apply automatically during maintenance windows. You can pin a specific major version if your application has known compatibility constraints.
Can I install Postgres extensions like PostGIS, pgvector, TimescaleDB?
Yes. Common extensions (PostGIS, pgvector, TimescaleDB, Citus, pg_cron, pg_stat_statements, uuid-ossp, pgcrypto) are pre-installed and available with CREATE EXTENSION. Custom extensions can be installed on Business+ plans.
How is connection pooling handled?
PgBouncer runs in front of every Postgres instance by default. Transaction-mode pooling handles most application workloads; session-mode is available for applications that use session-level features (advisory locks, prepared statements, temp tables). Application connects to PgBouncer's port; PgBouncer manages the backend connection pool.
What if I need vector search for AI/embeddings workloads?
pgvector is pre-installed. Create a vector column, add an HNSW or IVFFlat index, and query with the standard vector operators. Combined with pgvectorscale on Business+, Postgres handles vector search up to millions of embeddings without a separate vector database.
What happens if I want to migrate to RDS or another provider?
pg_dump / pg_dumpall for logical exports, or a WAL-based logical replication to your target for zero-downtime migration. Postgres is portable by design. Migration off ManageStacks is a supported operation — we help with the pg_dump or replication setup.
Deploy PostgreSQL in under 5 minutes.
Subscribe through your AWS, Azure, or GCP marketplace. We handle provisioning, SSL, monitoring, backups, updates, and security. From $29/app/month.