Size catalogue

Each running instance is billed per hour at its size price; the monthly column is that price x 730 hours. Disk is provisioned and billed separately per GB-hour above your plan’s allowance. Prices are USD list prices. Solo and Team include one instance: a monthly credit covers one instance of the smallest paid size, and every additional instance or larger size is billed at the full price (usage and billing). The capacity column is in compute units (1 CU = 1 vCPU + 4 GiB), used for capacity planning and usage charts only.

Instance sizes: vCPU, memory, storage ceiling, connection limit, capacity and compute price

Size	Tiers	vCPU	Memory	Max storage	Max connections	Capacity (CU)	$ / hour	≈ $ / month (730 h)
f0	free	0.063 (burst 1)	512 MiB	1 GB	20	0	Free	Free plan only
s0	standard, hipaa	0.25	1 GiB	20 GB	60	0.25	$0.0137	$10
s1	standard, hipaa	0.5	2 GiB	50 GB	100	0.5	$0.0205	$15
s2	standard, hipaa	1	4 GiB	200 GB	200	1	$0.0822	$60
m2	standard, hipaa	2	8 GiB	500 GB	400	2	$0.1507	$110
m4	standard, hipaa	4	16 GiB	1 TB	800	4	$0.2877	$210
l8	standard, hipaa	8	32 GiB	4 TB	1,500	8	$0.5616	$410
l16	standard, hipaa	16	64 GiB	8 TB	3,000	16	$1.3151	$960
xl32	standard, hipaa	32	128 GiB	16 TB	5,000	32	$2.5616	$1,870

• f0 runs only on free cells (Free plan only): burstable CPU, sleeps after 15 idle minutes, 1 GB of storage.
• Resize up or down between any two sizes of the same tier; the operator takes a backup first and the change is a rolling restart for operator-backed engines (see Resize).
• max_connections is enforced by the engine and by the gateway. The built-in pooler lets many short-lived clients share a smaller number of server connections (see Connection pooling).