Getting Started
Rate limits
How many requests each API key can make, what happens when you hit a limit, and how to stay under it.
Rate limits keep one busy integration from slowing down everything else. Each limit is set by your plan and counted per API key, so two services with separate keys never use up each other's allowance.
What's limited
| Traffic | Endpoints | Counted |
|---|---|---|
| Writes | POST /v1/dock/ingest | Per minute and per day, per key |
| Memory writes | POST /v1/dum/write | Against your plan's memory write limits |
| Reads | POST /v1/engine/retrieve, POST /v1/dum/search, POST /v1/dum/federated-search | Per minute and per day, per key |
| Answers | POST /v1/engine/query, POST /v1/engine/query/stream | Per key, and against your plan's monthly allowance |
Writes from the dashboard, such as uploads in Sources, don't count against your API key's write limit.
When you hit a limit
| Status | Meaning | What to do |
|---|---|---|
429 | A per-minute or per-day limit was hit. The message names which one. | Wait for the number of seconds in the Retry-After header, then retry. |
402 | Your plan's monthly allowance is used up. | Wait for the next cycle, or change plans. |
HTTP/1.1 429 Too Many Requests
Retry-After: 12Stay under the limits
- Use the SDK. It waits for
Retry-Afterand retries429for you. See Errors and retries. - Batch writes. One ingest can carry up to 100 events.
Dockbatches for you. - Use a key per service. Each key has its own allowance.
- Read once per turn. Recall and retrieve at the start of a turn, not on every token.
See your usage
Open Usage in the Docklee dashboard to see how much each runtime has used, and your plan's limits.
Next steps
You know the limits. Start with the endpoint every write goes through.