Limits & SLAs
Rate limits, storage quotas, retention, and Platform availability targets.
Overview
To guarantee stability, low latency, and high availability for all projects, sovseal enforces strict quotas and rate limits at the edge API.
Replication Rate Limits
Rate limits are applied per project API key. When a client exceeds limits, the server returns an HTTP 429 Too Many Requests code with a Retry-After header.
| Tier | Rate Limit (Requests per Minute) | Burst Allowance |
|---|---|---|
| Free | 60 RPM | 90 requests |
| Pro | 1,000 RPM | 1,500 requests |
| Team | 5,000 RPM | 7,500 requests |
| Enterprise | Custom | Custom |
Payload Size Limits
To optimize transmission speed and vector embedding processes:
- Maximum Ciphertext Payload:
5 MBperstore_memorycall. - Payload Rejection: Any write request exceeding
5 MBis blocked at the edge function and returns an HTTP413 Payload Too Largeerror.
Storage Quotas & Retention Windows
Storage limits apply to the cumulative size of all snapshots saved under a project.
Retention Windows
- Free Tier: Lineage history is retained for 30 days. Older snapshots are automatically pruned.
- Pro & Team Tiers: Lineage history is retained indefinitely or until manually cleared by the user.
Service Level Agreement (SLA) & Availability
We commit to a high-availability infrastructure for our Hosted Sync edge services.
Availability Target
- Pro & Team Tiers: 99.9% uptime SLA target.
- Enterprise: 99.95% uptime SLA target with financial guarantees.
Incident Communication
Our engineering team monitors edge health 24/7. System status, incident notifications, and scheduled maintenance announcements are posted on our status page at status.sovseal.com.