Skip to content

Add logging for tool execution and error handling in RequestInfoFromTableWithAIUseCaseV7 #1638

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Artuomka merged 1 commit into from
Feb 26, 2026
Merged

Add logging for tool execution and error handling in RequestInfoFromTableWithAIUseCaseV7 #1638

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
31 changes: 28 additions & 3 deletions backend/src/entities/ai/use-cases/request-info-from-table-with-ai-v7.use.case.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,5 @@
import { BaseMessage } from '@langchain/core/messages';
import { BadRequestException, Inject, Injectable, NotFoundException, Scope } from '@nestjs/common';
import { BadRequestException, Inject, Injectable, Logger, NotFoundException, Scope } from '@nestjs/common';
import { getDataAccessObject } from '@rocketadmin/shared-code/dist/src/data-access-layer/shared/create-data-access-object.js';
import { ConnectionTypesEnum } from '@rocketadmin/shared-code/dist/src/shared/enums/connection-types-enum.js';
import { IDataAccessObject } from '@rocketadmin/shared-code/dist/src/shared/interfaces/data-access-object.interface.js';
Expand Down Expand Up @@ -38,6 +38,7 @@ export class RequestInfoFromTableWithAIUseCaseV7
extends AbstractUseCase<RequestInfoFromTableDSV2, void>
implements IRequestInfoFromTableV2
{
private readonly logger = new Logger(RequestInfoFromTableWithAIUseCaseV7.name);
private readonly maxDepth: number = 10;
private readonly aiProvider: AIProviderType = AIProviderType.BEDROCK;

Expand Down Expand Up @@ -184,10 +185,21 @@ export class RequestInfoFromTableWithAIUseCaseV7
}
}

this.logger.log(
`Tool loop iteration ${depth + 1}: toolCalls=${pendingToolCalls.map((tc) => tc.name).join(', ') || 'none'}, ` +
`contentLength=${accumulatedContent.length}`,
);

if (pendingToolCalls.length === 0) {
break;
}

for (const toolCall of pendingToolCalls) {
this.logger.log(
`Tool call: ${toolCall.name}, arguments=${JSON.stringify(toolCall.arguments)}`,
Comment on lines +198 to +199

Copilot AI Feb 26, 2026

Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Logging full toolCall.arguments can leak sensitive data (e.g., raw SQL queries / Mongo pipelines, user-provided values) into application logs and may also create very large log lines. Consider logging only safe metadata (tool name/id, argument keys, and/or argument lengths) and/or truncating/redacting values, ideally at debug/verbose level behind a feature flag.

Suggested change
this.logger.log(
`Tool call: ${toolCall.name}, arguments=${JSON.stringify(toolCall.arguments)}`,
const argumentKeys =
toolCall.arguments && typeof toolCall.arguments === 'object'
? Object.keys(toolCall.arguments)
: [];
const argumentLength = JSON.stringify(toolCall.arguments ?? '').length;
this.logger.log(
`Tool call: ${toolCall.name}, argumentKeys=${argumentKeys.join(',')}, argumentLength=${argumentLength}`,

Copilot uses AI. Check for mistakes.
);
}

const toolResults = await this.executeToolCalls(
pendingToolCalls,
dataAccessObject,
Expand All @@ -196,6 +208,12 @@ export class RequestInfoFromTableWithAIUseCaseV7
foundConnection,
);

for (const toolResult of toolResults) {
this.logger.log(
`Tool result for ${toolResult.toolCallId}: resultLength=${toolResult.result.length}`,
);
}

if (this.aiProvider === AIProviderType.OPENAI && lastResponseId) {
currentConfig = { ...currentConfig, previousResponseId: lastResponseId };
}
Expand All @@ -209,11 +227,15 @@ export class RequestInfoFromTableWithAIUseCaseV7

depth++;
} catch (loopError) {
this.logger.error(
`Error in tool loop at depth ${depth + 1}: ${loopError.message}`,
);
throw loopError;
}
}

if (depth >= this.maxDepth) {
this.logger.warn(`Tool loop reached max depth (${this.maxDepth})`);
const maxDepthMessage =
'\n\nYour question is too complex to process at this time. Please try simplifying it or breaking it down into smaller parts.';
response.write(maxDepthMessage);
Expand Down Expand Up @@ -284,6 +306,9 @@ export class RequestInfoFromTableWithAIUseCaseV7
result = encodeError({ error: `Unknown tool: ${toolCall.name}` });
}
} catch (error) {
this.logger.error(
`Tool call ${toolCall.name} (${toolCall.id}) failed: ${error.message}`,
);
result = encodeError({ error: error.message });
}

Expand Down Expand Up @@ -409,8 +434,8 @@ export class RequestInfoFromTableWithAIUseCaseV7

private async generateAndUpdateChatName(chatId: string, userMessage: string): Promise<void> {
try {
const CHAT_NAME_GENERATION_PROMPT = `Generate a very short, concise title (max 5-6 words) for a chat conversation based on the user's first question.
The title should capture the main topic or intent.
const CHAT_NAME_GENERATION_PROMPT = `Generate a very short, concise title (max 5-6 words) for a chat conversation based on the user's first question.
The title should capture the main topic or intent.
Respond ONLY with the title, no quotes, no explanation.
User question: `;
const prompt = CHAT_NAME_GENERATION_PROMPT + userMessage;
Expand Down
Loading