Why Relay?
How Relay prices managed PostgreSQL — machine tiers, a platform fee that credits toward usage, and the governance layer included with every plan.
A Relay cluster is a managed PostgreSQL cluster plus the governance layer — masking, a query firewall, access grants, and a full audit trail — at one price. Every tier includes storage, high availability is billed as node count rather than as a premium, and the fee you pay is a floor that your usage draws down. This page explains how that adds up.
Prices on this page are illustrative USD list rates and may change. Your organization's live rates are always under Organization → Billing. Organizations in Brazil can pay in reais — see Billing in Brazilian reais.
How Relay bills
- A platform fee that credits toward usage. Team is $50/month and Business is $500/month. The fee is a floor, not a surcharge — usage draws down against it first. See Platform fee.
- Compute by node-hours. Each running node accrues hours at its machine tier rate; a full month is 730 hours.
- High availability is node count. A primary plus a replica bills as two nodes of the tier. There is no separate HA SKU or multiplier.
- Storage included per tier. Every tier ships with a storage allowance; provisioned storage beyond it bills at $0.30 per GiB-month.
- Pause stops compute. A paused cluster accrues no node-hours; storage and backups are kept.
Backups, point-in-time recovery, and database metrics are part of the cluster, not separate line items. See Usage & metering.
Machine tiers
Per node, per month, in USD.
| Tier | vCPU | Memory | Included storage | Price |
|---|---|---|---|---|
| R5 | 0.5 | 1 GiB | 5 GiB | $0 (free plan) |
| R10 | 1 | 2 GiB | 10 GiB | $25 |
| R20 | 2 | 4 GiB | 20 GiB | $49 |
| R30 | 2 | 6.25 GiB | 40 GiB | $105 |
| R40 | 4 | 13.75 GiB | 80 GiB | $215 |
| R50 | 8 | 29 GiB | 160 GiB | $425 |
| R60 | 16 | 59.75 GiB | 320 GiB | $850 |
| R80 | 32 | 121.5 GiB | 750 GiB | $1,725 |
| R100 | 48 | 183 GiB | 1000 GiB | $2,575 |
| R200 | 64 | 244.5 GiB | 1500 GiB | $3,450 |
R5, R10 and R20 are burstable: they run on shared nodes, are guaranteed a third of their vCPU, and can use the full vCPU when the node has spare capacity. R30 and larger get their whole vCPU. Memory is always reserved in full.
Tiers can be resized at any time (a resize restarts the nodes, so expect a brief interruption); a cluster's bill follows the hours it spends on each tier.
What's included in the price
A tier's price is the whole managed database, not a compute line with extras bolted on:
- Storage is included. Every tier ships with storage sized for the tier; you only pay for what you provision beyond it.
- High availability is just node count. A primary plus a replica bills as two nodes — there is no HA premium or separate storage multiplier.
- No per-I/O charges. There is no I/O meter; your bill doesn't move with query volume.
- Backups and PITR are part of the cluster. Schedules, retention tiers, and point-in-time recovery are not separate line items. See Backups & restore.
- Metrics are included on Team and above. See Metrics & logs.
- Pause when idle. A paused cluster accrues no compute; you keep storage and backups.
- Governance is bundled. Masking, a query firewall, and access grants come with Team and above, and every plan keeps a per-person audit trail — none of them are add-ons.