Convex
Forward logs from Convex functions to Sazabi for AI-powered analysis and investigation.
Convex is a native data source. Sazabi connects to your Convex projects to receive function logs and events automatically.
About this data source
When you connect your Convex team, Sazabi configures log streaming to receive logs from your Convex functions in real-time. This gives you:
- Automatic log forwarding without managing infrastructure
- Real-time streaming of function logs and events
- Selective forwarding from specific projects
- AI-powered analysis and investigation of your backend logs
Prerequisites
Before you begin, make sure you have:
- A Convex account with a Pro plan (log streaming requires a paid plan)
- A team access token from the Convex dashboard
- Projects you want to forward logs from
Log streaming is only available on Convex Pro plans. If you are on the Free plan, you will need to upgrade before setting up this integration.
Setup
Get your team access token
Generate a team access token in your Convex dashboard:
- Go to your Convex Dashboard
- Navigate to your team settings
- Find the Access Tokens section
- Click Generate Token
- Copy the token and store it securely
Team access tokens provide access to all projects in your team. Keep this token secure and do not share it.
Connect in Sazabi
Select Convex from the list of available data sources.
Enter your team access token when prompted.
Select projects
After authenticating, Sazabi displays a list of your Convex projects. Select the projects you want to forward logs from.
Each project has separate development and production deployments. You can choose to forward logs from one or both environments.
Confirm and connect
Review your selections and click Connect. Sazabi will configure log streaming for each selected project. Logs will begin flowing within a few minutes.
What Sazabi creates
When you connect Convex, Sazabi configures the following:
- Log stream subscription for each selected project and deployment
Log streams are managed through the Convex platform. Unlike other native data sources, Sazabi does not create additional infrastructure in your Convex account.
Log types
Convex log streaming forwards the following:
- Function logs: Output from your Convex functions (queries, mutations, actions)
- Scheduled function logs: Logs from scheduled and cron jobs
- HTTP action logs: Logs from HTTP endpoints
- System events: Internal Convex events and errors
All log types are automatically tagged in Sazabi, allowing you to filter by function type and deployment environment.
Verifying logs are flowing
Once connected, you can verify that logs are flowing to Sazabi:
-
Ask the assistant: Open a thread in Sazabi and ask "Show me recent Convex logs" or "What errors appeared in my mutations today?"
-
Check the dashboard: Navigate to your project in the Sazabi dashboard to see logs appear in real-time.
-
Trigger a function: Call a Convex function from your application and verify the logs appear in Sazabi.
Disconnecting
When you disconnect Convex from Sazabi:
- Sazabi removes all log stream subscriptions it configured
- No logs are deleted from Convex
- Historical logs already sent to Sazabi remain available until retention expires
To disconnect, go to Settings > Data Sources, find your Convex connection, and click Disconnect.
Troubleshooting
Pro plan required
Log streaming is only available on Convex Pro plans. If you see an error about log streaming not being available, upgrade your team to Pro in the Convex dashboard.
Token expired
Team access tokens can expire or be revoked. If the connection stops working, generate a new token in the Convex dashboard and update the connection in Sazabi.
Project not found
If a project does not appear in the list, verify it belongs to the team associated with your access token. Projects in other teams require separate connections.
Logs not arriving
- Verify the log stream is configured in your Convex project
- Trigger a function call to generate fresh logs
- Check that your Sazabi API key is valid and the project is active
- Allow a few minutes for logs to propagate through the pipeline