Operate

Usage & Limits

Rate limits, quotas, tracked metrics, and what happens when limits are reached.

Tracked Metrics

The console dashboard tracks usage per project in real time:

Metric	Description
Read operations	Total search and query calls
Write operations	Total insert, update, and delete calls
Vectors stored	Current count of vectors in your project
Storage size	Estimated disk footprint of your data
Daily usage history	Read/write trends over the past 30 days

Rate Limits

The gateway enforces per-project rate limits to ensure fair resource sharing on the managed platform:

Read and write traffic is rate limited independently.
API key creation and management calls have separate, lower limits.
Rate limits apply per project, not per key — all keys in a project share the same quota.

Quotas

Service-level quotas are enforced for:

Projects per account — Maximum number of projects you can create.
API keys per project — Maximum active keys per project.
Vectors per project — Maximum stored vectors (varies by tier).
Storage per project — Maximum disk usage (varies by tier).

Limits May Change

KyroDB is in early access. Specific numeric limits are set conservatively and will be adjusted as the platform scales. Contact support if you need higher limits for your use case.

What Happens When Limits Are Reached

Status Code	Cause	What To Do
`429`	Rate limit exceeded for the current window	Retry with exponential backoff
`403`	Quota exceeded (e.g., max vectors reached)	Delete unused data or upgrade your tier
`400`	Request exceeds batch size limit	Reduce batch size and retry

Best Practices

Implement exponential backoff for 429 responses.
Use x-idempotency-key headers on write requests so retries are safe.
Monitor your usage in the console dashboard to avoid hitting quotas unexpectedly.
Use namespaces to organize data instead of creating extra projects.