Instance Types

PostgreSQL Instance Types

PostgreSQL on Thalassa Cloud DBaaS runs on General Purpose instance types. Each type has a fixed vCPU and memory size; the platform sets PostgreSQL parameters to values suited for that size. You can override allowed parameters if your workload needs different tuning.

General Purpose (db-pgp)

General purpose instance types for production PostgreSQL workloads.

Instance type	vCPUs	Memory	Description
micro	1	1 GB	Micro instance for light workloads, development, or testing.
small	1	2 GB	Small instance for low-traffic or small datasets.
medium	2	4 GB	Medium instance for typical application workloads.
large	2	8 GB	Large instance for production with higher memory needs.
xlarge	4	16 GB	Extra large instance for heavier production workloads.
2xlarge	8	32 GB	Double extra large for high concurrency and large datasets.

How instance types affect PostgreSQL

When you select an instance type, Thalassa Cloud applies default parameters for that size:

Memory-related — work_mem, temp_buffers, maintenance_work_mem, effective_cache_size, and WAL-related settings scale with instance size.
Parallelism — max_worker_processes, max_parallel_workers, max_parallel_workers_per_gather, and max_parallel_maintenance_workers increase on larger instances.
Connections — max_connections is set per size (e.g. 100 on smaller instances, 200 on xlarge and 2xlarge).

You can change any allowed parameter after creation to better match your workload. Instance-type defaults are a starting point, not a lock-in.

Default parameters per instance type

The table below lists the default PostgreSQL parameters applied for each General Purpose instance type. These values are tuned for that instance size; you can override them within the allowed parameters.

Parameter	micro	small	medium	large	xlarge	2xlarge
`max_connections`	100	100	100	100	200	200
`work_mem`	4 MB	4 MB	8 MB	16 MB	32 MB	64 MB
`temp_buffers`	8 MB	16 MB	32 MB	64 MB	128 MB	256 MB
`maintenance_work_mem`	256 MB	512 MB	1 GB	1 GB	2 GB	2 GB
`effective_cache_size`	512 MB	1 GB	2 GB	4 GB	10 GB	20 GB
`checkpoint_completion_target`	0.9	0.9	0.9	0.9	0.9	0.9
`wal_buffers`	16 MB	16 MB	16 MB	16 MB	16 MB	16 MB
`min_wal_size`	512 MB	1 GB	1 GB	1 GB	2 GB	4 GB
`max_wal_size`	2 GB	4 GB	6 GB	6 GB	10 GB	16 GB
`random_page_cost`	1.2	1.2	1.2	1.2	1.2	1.2
`effective_io_concurrency`	200	200	200	200	200	200
`max_worker_processes`	1	1	4	4	8	12
`max_parallel_workers_per_gather`	0	0	2	2	4	8
`max_parallel_workers`	1	1	4	4	8	12
`max_replication_slots`	4	4	5	5	6	9
`max_parallel_maintenance_workers`	2	0	2	2	4	8
`statement_timeout`	0	0	0	0	0	0

statement_timeout value 0 means disabled (no timeout). Other parameters may have cluster-level defaults in addition to these; see Parameters for the full list of configurable parameters.

Choosing an instance type

Development / testing — micro or small.
Production (moderate traffic) — medium or large.
Production (high traffic, many connections, or large working set) — xlarge or 2xlarge.

If you hit connection limits, high I/O wait, or slow queries despite tuning, consider moving to a larger instance type.

Parameters