This document describes the default configuration of Conversation Learner, and key service boundaries.
| Parameter | Default value |
|---|---|
| Default session timeout | 30 minutes |
| Parameter | Limit |
|---|---|
| Authoring API, max HTTP calls per month | 5M |
| Authoring API, max HTTP calls per second | 25 |
| Session API, max HTTP calls per month | 500K |
| Session API, max HTTP calls per second | 10 |
| Max number of custom (non-programmatic) entities per model | See LUIS boundaries doc; in practice, actual number may be slightly smaller |
| Max number of pre-built entities per model | See LUIS boundaries doc |
| Max number of entities (in total) per model | 100 |
| Max number of actions per model | 32 |
| Max number of train dialogs per model | 1000 |
| Max number of user turns per train dialog | 100 |
| Max number of log dialogs per model | No pre-set limit, but log dialogs are only retained for a fixed period before being discarded. Also, the Conversation Learner UI will show 100 log dialogs at a time. |
| Max number of models per user | No pre-set limit |
| Max number of sequential non-wait actions | 5 (*) |
(*) After 5 sequential non-wait actions, all non-wait actions are masked, and Conversation Learner will choose amongst available wait actions.
[!div class="nextstepaction"] Get started with Conversation Learner