Configure global limits, audit retention, and optional performance features for your entire Odoo MCP Connection.
Overview
System Settings control how the MCP connector behaves for your whole Odoo database. Individual API keys can override some limits (like rate limits), but these settings provide the defaults and safety boundaries.
Open them from Odoo MCP Connection → Configuration → Settings. Changes apply after you click Save.
There is also a built-in Help page under Configuration with technical endpoint details for IT teams.
How It Works
Settings are stored as system parameters and read every time an AI request arrives. For example:
- If Enable MCP Connection is off, all AI requests are rejected immediately.
- If a key has no custom rate limit, the Default Rate Limits apply.
- A scheduled task deletes audit logs older than Audit Log Retention days.
- If Redis is enabled, rate limiting and schema caching are shared across multiple Odoo workers.
Step-by-Step Guide
Go to Odoo MCP Connection → Configuration → Settings.
Under MCP Endpoint, confirm Enable MCP Connection is checked.
Under Rate Limits & Retention, set defaults that match your server capacity.
Set Audit Log Retention based on how long you need compliance records.
Set Max Active API Keys if you want to cap how many keys can exist (0 = unlimited).
Under Redis (optional), enable only if your IT team has set up a Redis server.
Click Save at the top of the settings page.
Test an API key to confirm settings work as expected.
Fields Table
| Field Name | Description | Example |
|---|---|---|
Enable MCP Connection |
Master switch – turn off to block all AI access |
On |
Default Rate Limit (per second) |
Max AI requests per second per key (unless key overrides) |
10 |
Default Rate Limit (per minute) |
Max AI requests per minute per key |
120 |
Max Search Limit |
Maximum records returned in a single search |
200 |
Audit Log Retention (days) |
How long to keep audit logs before auto-deletion |
7 |
Max Active API Keys |
Maximum number of active keys allowed (0 = no limit) |
0 |
Schema Cache TTL (seconds) |
How long to cache data structure info for performance |
300 |
Audit Payload Max Chars |
Maximum characters stored per request/response in logs |
8000 |
Enable Redis |
Use Redis for shared rate limiting and caching |
Off |
Redis URL |
Connection address for your Redis server |
redis://127.0.0.1:6379/0 |
Field Explanations
Enable MCP Connection
The emergency off switch. Use during maintenance or security incidents to stop all AI traffic instantly without deleting keys.
Default Rate Limit (per second & per minute)
Protects your server from too many rapid AI requests. Keys with their own limits (non-zero values) override these defaults.
Max Search Limit
Caps how many records AI can fetch in one search. Prevents accidental “download entire database” requests. Default 200 is reasonable for most businesses.
Audit Log Retention (days)
Balances compliance needs with database size. Increase for regulated industries; decrease to save storage on busy systems.
Max Active API Keys
Optional governance control. Set to 10, for example, to force teams to archive unused keys before creating new ones. Zero means unlimited.
Schema Cache TTL (seconds)
Technical performance setting. Caches information about Odoo data structures. Higher values = faster but slightly less fresh. Default 300 (5 minutes) is fine for most setups.
Audit Payload Max Chars
Limits how much request/response text is stored per log entry. Increase if you need more detail in audit reviews; decrease to save database space.
Enable Redis & Redis URL
Only needed for multi-server Odoo deployments where rate limits must be shared across workers. Single-server setups can leave this off. Ask your IT team before enabling.
Tips
- Start with default values and adjust only if you see rate limit errors in audit logs or performance issues.
- For compliance (GDPR, SOX), set audit retention to match your company’s data retention policy.
- The built-in Help page (Configuration → Help) shows endpoint URLs and tool lists – share it with your IT team.
Common Mistakes
- Leaving MCP disabled after maintenance: AI tools will get errors until you re-enable the connection. Add this to your maintenance checklist.
- Setting rate limits too low: Busy AI workflows may hit limits quickly. Monitor audit logs for rate limit errors and adjust upward.
- Enabling Redis without IT setup: Redis requires a running server at the URL you specify. Enabling it without proper setup can cause connection errors.