Export Lovable logs to Azure Log Analytics

A data export to Azure Log Analytics sends rows to a table in your Log Analytics workspace. Each source gets its own table, which a setup script creates together with a data collection rule. Lovable signs in as an app registration in your Microsoft Entra tenant through a federated credential, so no secret is stored on either side. For sources, schedules, and run behavior, see Data exports.

Prerequisites

A Log Analytics workspace in the Azure public cloud. Azure Government, Azure China, and private endpoints are not supported.
Permission to register an app and add a federated credential in Microsoft Entra ID.
An Azure account that can create tables, data collection rules, custom roles, and role assignments in the workspace’s resource group.
The Azure CLI, signed in to the workspace’s subscription, in a bash or zsh shell.
The workspace admin or owner role in Lovable, on an Enterprise plan.

Set up the export

Set up one export per source. Run the setup script shown in each export’s setup panel, because each source has its own table and rule.

Start a new export

Go to Data exports, click New data export, and choose Azure Log Analytics under Destination. Choose the Source now. The setup script depends on it.

Register one app for this Lovable workspace. In the Azure portal, open Microsoft Entra ID → App registrations → New registration. On the app’s Overview page, copy the Directory (tenant) ID into Tenant ID and the Application (client) ID into Client ID. You can also run the CLI commands shown in the setup panel.Entra allows one federated credential per issuer and subject on an app. A second Lovable workspace needs a second app.

Trust Lovable

On the app, open Certificates & secrets → Federated credentials → Add credential. Choose the scenario Other issuer and fill in the Credential values from the setup panel. You can also run the CLI commands shown in the setup panel.

Enter the workspace resource ID

In Azure, open the Log Analytics workspace’s Properties page and copy its Resource ID. Paste it into Log Analytics workspace resource ID. It looks like

/subscriptions/<subscription ID>/resourceGroups/<resource group>/providers/Microsoft.OperationalInsights/workspaces/<workspace name>

Create the table and grant access

Copy the Azure CLI script from the setup panel and run it. The script is ready to paste and already holds your values. It creates or updates the table for your source and its data collection rule, then lets the app read the workspace, that table, and that rule, and send data through the rule.You can run the script again at any time. Role assignments can take several minutes to apply.

Test the connection

Click Test connection. Access check succeeded means that Azure authentication, the table and rule, and the permissions on them passed. The test sends no events. The first run confirms that ingestion works.

Create the export

Click Create export.

The setup panel is the source of truth for the federated credential and the permissions. Run the script shown in the setup panel rather than creating the roles yourself.

What Lovable can access

The app that Lovable signs in as can read the Log Analytics workspace and the tables and data collection rules that the setup script created for it, check its own permissions on them, and send data through those rules. Lovable cannot change any table or rule. It has no write, delete, role assignment, workspace key, or log query access.

Tables and columns

Each source has its own table:

Source	Table
Audit logs	`LovableAuditLogsV2_CL`

Lovable sends each source row in the dynamic Payload column, using the field names and JSON types listed in Sources. Each table has these columns:

Column	Contents
`TimeGenerated`	When the event happened, in UTC (the source’s `timestamp`).
`EventTime`	When the event happened, in UTC (the source’s `timestamp`).
`EventId`	The row’s `id`. Use it to remove duplicates.
`WorkspaceId`	ID of the Lovable workspace.
`ProjectId`	Empty for audit logs.
`LovableExportId`	ID of the export.
`LovableRunId`	ID of the run that delivered the row.
`Payload`	The source row as a JSON object, including nested values. For the fields, see Sources. Azure size limits apply.

Rows appear in the table after Azure processes them. The first delivery after setup can take several minutes. The tables use the Analytics plan and the retention of your workspace. New source fields appear inside Payload without a table or data collection rule update.

Query and remove duplicates

Delivery is at least once, so a retried run can send a row again. Remove duplicates in queries with EventId:

LovableAuditLogsV2_CL
| summarize arg_max(TimeGenerated, *) by EventId

Audit log Payload.details contains JSON text. Parse it to query its properties:

LovableAuditLogsV2_CL
| extend Details = parse_json(tostring(Payload.details))
| project TimeGenerated, Action = tostring(Payload.action), Details

Change or remove the export

You cannot change the tenant ID, the client ID, or the workspace resource ID of an export. To change them, create a new export. Deleting an export keeps its table, its data collection rule, and its data in Azure. To remove them, disable or delete the export first, then delete them in Azure. Lovable cannot recreate a table or rule that you delete. Runs fail until you run the setup script again.

Azure size limits

Azure limits each Logs Ingestion API request to 1 MB, including event payloads and the metadata columns Lovable adds to each row (EventId, WorkspaceId, LovableExportId, LovableRunId, and the time columns). This request limit does not guarantee that individual fields are retained in full. See Azure Monitor service limits.

Azure can accept a request and still shorten field values longer than 64 KB. A run that Azure accepted counts as delivered, even if Azure shortened some values. Lovable sends every value in full and does not read it back.

Troubleshooting

Authentication denied

Check the federated credential’s issuer, subject, and audience against the setup panel. Then run the setup script again so the role assignments exist. Role assignments can take several minutes to apply.

Azure resource conflict

A table or rule with Lovable’s name exists, but the setup script did not create it, or it sends data to another workspace. Check the workspace resource ID. Or remove the old table or rule and run the setup script again.

Azure schema does not match

The table or rule does not match the export’s schema, for example after a column type or rule transformation was changed in Azure. Run the source’s setup script again. If a column has a different type than Lovable expects, restore that type first. Your existing data is not deleted.

Destination configuration is invalid

Check the tenant ID and the destination settings against the setup panel. If the source’s table or rule is missing, run the source’s setup script again. New permissions and resources can take several minutes to become available.

Azure Monitor rejected the request

Check the workspace resource ID, and that the table and rule match the setup panel.

A run fails because the data is too large

Azure Monitor rejects requests larger than its 1 MB limit, including Lovable metadata. If a single row cannot fit, the run fails. Contact Lovable support. Requests within this limit can still have oversized fields truncated, as described in Azure size limits.

See Troubleshooting for general run issues.

Getting started

Build

Connect

Collaborate

Deploy & host

Grow & optimize

Workspace & account

Plans & billing

Guides & reference

Export Lovable logs to Azure Log Analytics

Prerequisites

Set up the export

What Lovable can access

Tables and columns

Query and remove duplicates

Change or remove the export

Azure size limits

Troubleshooting

​Prerequisites

​Set up the export

​What Lovable can access

​Tables and columns

​Query and remove duplicates

​Change or remove the export

​Azure size limits

​Troubleshooting

Prerequisites

Set up the export

What Lovable can access

Tables and columns

Query and remove duplicates

Change or remove the export

Azure size limits

Troubleshooting