Include deploymentId in Langfuse telemetry metadata#2828
Open
attadevelops wants to merge 1 commit intogiselles-ai:mainfrom
Open
Include deploymentId in Langfuse telemetry metadata#2828attadevelops wants to merge 1 commit intogiselles-ai:mainfrom
attadevelops wants to merge 1 commit intogiselles-ai:mainfrom
Conversation
|
Someone is attempting to deploy a commit to the Giselle Team on Vercel. A member of the Team first needs to authorize it. |
|
|
Finished running flow.
|
||||||||||||||||||
Contributor
📝 WalkthroughWalkthroughThe trace function is simplified by removing team-specific context and consolidating the tracing logic. A global Langfuse client is initialized, and the new trace function passes essential parameters (traceId, userId, sessionId, metadata) along with a deploymentId derived from environment variables to a streamlined traceGeneration function. Changes
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~30 minutes Poem
🚥 Pre-merge checks | ✅ 3 | ❌ 2❌ Failed checks (1 warning, 1 inconclusive)
✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
📝 Coding Plan
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment Tip CodeRabbit can generate a title for your PR based on the changes.Add |
Review Summary by QodoRefactor trace generation and add deploymentId metadata
WalkthroughsDescription• Refactored trace generation to simplified, generic implementation • Added deploymentId field to Langfuse trace metadata • Moved from complex generation-specific logic to lightweight wrapper • Simplified trace.ts to use new generic traceGeneration function Diagramflowchart LR
A["trace.ts<br/>traceGenerationForTeam"] -->|"refactored to"| B["trace.ts<br/>trace function"]
B -->|"calls"| C["trace-generation.ts<br/>traceGeneration"]
C -->|"creates trace with"| D["Langfuse<br/>metadata + deploymentId"]
E["VERCEL_DEPLOYMENT_ID<br/>env var"] -->|"passed as"| D
File Changes1. apps/studio.giselles.ai/lib/trace.ts
|
Code Review by Qodo
1. trace.ts imports not organized
|
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: be22e60413
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
| import type { ModelMessage, ProviderMetadata } from "ai"; | ||
| import type { CurrentTeam } from "@/services/teams"; | ||
| import { Langfuse } from "langfuse"; | ||
| import { traceGeneration } from "@repo/langfuse/trace-generation"; |
There was a problem hiding this comment.
This import switches to @repo/langfuse/trace-generation, but apps/studio.giselles.ai/tsconfig.json only maps @/* and @giselles-ai/*, so the module cannot be resolved during type-check/build. In this state the studio app fails to compile before telemetry runs, so the deploymentId change never reaches production.
Useful? React with 👍 / 👎.
| generation: CompletedGeneration | FailedGeneration; | ||
| inputMessages: ModelMessage[]; | ||
| outputFileBlobs?: OutputFileBlob[]; | ||
| export const trace = ({ |
There was a problem hiding this comment.
Replacing traceGenerationForTeam with trace breaks existing callback wiring, because callers in apps/studio.giselles.ai/app/giselle.ts and apps/studio.giselles.ai/trigger/generate-content-job.ts still import and invoke traceGenerationForTeam. That missing export causes a hard compile/runtime-module error for generation callbacks, so telemetry code paths cannot execute.
Useful? React with 👍 / 👎.
![qodo-free-for-open-source-projects[bot]](https://avatars.githubusercontent.com/in/1068000?s=60&v=4){kind=small}
Comment on lines
+1
to
+2
| import { Langfuse } from "langfuse"; | ||
| import { traceGeneration } from "@repo/langfuse/trace-generation"; |
There was a problem hiding this comment.
1. trace.ts imports not organized 📘 Rule violation ✓ Correctness
The new imports in apps/studio.giselles.ai/lib/trace.ts are not organized/sorted in a Biome-compatible way, which can cause formatting/CI failures and inconsistent style. Biome typically expects organized imports (including consistent ordering).
Agent Prompt
## Issue description
`apps/studio.giselles.ai/lib/trace.ts` has newly-added imports that are not organized/sorted in a Biome-compatible way.
## Issue Context
PR Compliance requires Biome-compatible formatting, including organized imports.
## Fix Focus Areas
- apps/studio.giselles.ai/lib/trace.ts[1-2]
ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools
Comment on lines
+10
to
+29
| export const trace = ({ | ||
| traceId, | ||
| userId, | ||
| sessionId, | ||
| metadata, | ||
| }: { | ||
| traceId: string; | ||
| userId?: string; | ||
| sessionId?: string; | ||
| userId: string; | ||
| team: TeamForPlan; | ||
| providerMetadata?: ProviderMetadata; | ||
| requestId?: string; | ||
| }) { | ||
| const teamPlan = args.team.plan; | ||
| const planTag = `plan:${teamPlan}`; | ||
|
|
||
| await traceGeneration({ | ||
| generation: args.generation, | ||
| outputFileBlobs: args.outputFileBlobs, | ||
| inputMessages: args.inputMessages, | ||
| userId: args.userId, | ||
| tags: [planTag], | ||
| metadata: { | ||
| generationId: args.generation.id, | ||
| teamPlan, | ||
| userId: args.userId, | ||
| subscriptionId: args.team.activeSubscriptionId ?? "", | ||
| customerId: args.team.activeCustomerId ?? "", | ||
| providerMetadata: args.providerMetadata, | ||
| requestId: args.requestId, | ||
| workspaceId: args.generation.context.origin.workspaceId, | ||
| }, | ||
| sessionId: args.sessionId, | ||
| metadata?: Record<string, unknown>; | ||
| }) => { | ||
| return traceGeneration({ | ||
| langfuse, | ||
| traceId, | ||
| userId, | ||
| sessionId, | ||
| metadata, | ||
| deploymentId: process.env.VERCEL_DEPLOYMENT_ID, | ||
| }); | ||
| } | ||
| }; |
There was a problem hiding this comment.
2. Missing tracing export 🐞 Bug ✓ Correctness
apps/studio.giselles.ai/lib/trace.ts no longer exports traceGenerationForTeam, but studio runtime code still imports and calls it, causing a compile-time error. This blocks all code paths that trace generationComplete/generationError and trigger jobs.
Agent Prompt
## Issue description
`traceGenerationForTeam` was removed/replaced by `trace`, but existing studio code still imports and calls `traceGenerationForTeam`, causing a build failure.
## Issue Context
This impacts generation telemetry in both request-driven flows (`app/giselle.ts`) and Trigger jobs (`trigger/generate-content-job.ts`).
## Fix Focus Areas
- apps/studio.giselles.ai/lib/trace.ts[1-29]
- apps/studio.giselles.ai/app/giselle.ts[19-281]
- apps/studio.giselles.ai/trigger/generate-content-job.ts[1-103]
ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools
Comment on lines
+4
to
+8
| const langfuse = new Langfuse({ | ||
| secretKey: process.env.LANGFUSE_SECRET_KEY, | ||
| publicKey: process.env.LANGFUSE_PUBLIC_KEY, | ||
| baseUrl: process.env.LANGFUSE_BASE_URL, | ||
| }); |
There was a problem hiding this comment.
3. Wrong langfuse base env 🐞 Bug ⛯ Reliability
apps/studio.giselles.ai/lib/trace.ts reads LANGFUSE_BASE_URL, but the repo documents LANGFUSE_BASEURL, so baseUrl configuration will not be applied. This can route telemetry to the wrong Langfuse endpoint or fail in self-hosted setups.
Agent Prompt
## Issue description
The code reads `LANGFUSE_BASE_URL` but the repo’s documented env var is `LANGFUSE_BASEURL`, so `baseUrl` will stay undefined for correctly configured environments.
## Issue Context
This silently breaks base URL configuration for Langfuse.
## Fix Focus Areas
- apps/studio.giselles.ai/lib/trace.ts[4-8]
- apps/studio.giselles.ai/.env.example[22-26]
ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools
Comment on lines
+13
to
+32
| export const traceGeneration = ({ | ||
| langfuse, | ||
| traceId, | ||
| userId, | ||
| sessionId, | ||
| metadata, | ||
| deploymentId, | ||
| }: TraceGenerationParams): LangfuseTraceClient => { | ||
| const trace = langfuse.trace({ | ||
| name: "generation", | ||
| userId: args.userId ? String(args.userId) : undefined, | ||
| input: langfuseInput, | ||
| sessionId: args.sessionId, | ||
| }); | ||
|
|
||
| const tags = [...(args.tags ?? []), ...extractTags(args.operationNode)]; | ||
| const metadata = { | ||
| ...args.metadata, | ||
| ...extractMetadata(args.operationNode), | ||
| }; | ||
|
|
||
| if (args.generation.status === "failed") { | ||
| trace.update({ | ||
| tags, | ||
| metadata, | ||
| }); | ||
|
|
||
| trace.generation({ | ||
| name: "generateContent", | ||
| model: args.operationNode.content.languageModel.id, | ||
| modelParameters: args.operationNode.content.languageModel.configuration, | ||
| input: langfuseInput, | ||
| startTime: new Date(args.generation.startedAt), | ||
| endTime: new Date(args.generation.failedAt), | ||
| metadata, | ||
| level: "ERROR", | ||
| statusMessage: args.generation.error.message, | ||
| }); | ||
| await langfuse.flushAsync(); | ||
| return; | ||
| } | ||
|
|
||
| const usage = args.generation.usage ?? { | ||
| inputTokens: 0, | ||
| outputTokens: 0, | ||
| totalTokens: 0, | ||
| }; | ||
|
|
||
| const cost = await calculateDisplayCost( | ||
| args.operationNode.content.languageModel.provider, | ||
| args.operationNode.content.languageModel.id, | ||
| { | ||
| inputTokens: usage.inputTokens ?? 0, | ||
| outputTokens: usage.outputTokens ?? 0, | ||
| }, | ||
| ); | ||
|
|
||
| trace.update({ | ||
| output: args.generation.outputs, | ||
| tags, | ||
| metadata, | ||
| }); | ||
|
|
||
| trace.generation({ | ||
| name: "generateContent", | ||
| model: args.operationNode.content.languageModel.id, | ||
| modelParameters: args.operationNode.content.languageModel.configuration, | ||
| input: langfuseInput, | ||
| output: args.generation.outputs, | ||
| usage: { | ||
| unit: "TOKENS", | ||
| input: usage.inputTokens ?? 0, | ||
| output: usage.outputTokens ?? 0, | ||
| total: usage.totalTokens ?? 0, | ||
| inputCost: cost.inputCostForDisplay, | ||
| outputCost: cost.outputCostForDisplay, | ||
| totalCost: cost.totalCostForDisplay, | ||
| id: traceId, | ||
| userId, | ||
| sessionId, | ||
| metadata: { | ||
| ...metadata, | ||
| ...(deploymentId ? { deploymentId } : {}), | ||
| }, | ||
| startTime: new Date(args.generation.startedAt), | ||
| endTime: new Date(args.generation.completedAt), | ||
| metadata, | ||
| level: "DEFAULT", | ||
| }); | ||
|
|
||
| await langfuse.flushAsync(); | ||
| } | ||
|
|
||
| export async function traceGeneration(args: { | ||
| generation: CompletedGeneration | FailedGeneration; | ||
| inputMessages: ModelMessage[]; | ||
| userId?: string; | ||
| metadata?: Record<string, unknown>; | ||
| tags?: string[]; | ||
| outputFileBlobs?: OutputFileBlob[]; | ||
| sessionId?: string; | ||
| }) { | ||
| try { | ||
| const { operationNode } = args.generation.context; | ||
|
|
||
| // Skip telemetry for query and action nodes | ||
| if (isQueryNode(operationNode) || isActionNode(operationNode)) { | ||
| return; | ||
| } | ||
|
|
||
| // Handle content generation nodes separately | ||
| if (isContentGenerationNode(operationNode)) { | ||
| await traceContentGeneration({ | ||
| generation: args.generation, | ||
| operationNode, | ||
| inputMessages: args.inputMessages, | ||
| userId: args.userId, | ||
| metadata: args.metadata, | ||
| tags: args.tags, | ||
| sessionId: args.sessionId, | ||
| }); | ||
| return; | ||
| } | ||
|
|
||
| // Validate supported node types | ||
| if ( | ||
| !isTextGenerationNode(operationNode) && | ||
| !isImageGenerationNode(operationNode) | ||
| ) { | ||
| console.warn( | ||
| `Telemetry: Unsupported node type: ${operationNode.content.type}`, | ||
| ); | ||
| return; | ||
| } | ||
|
|
||
| // Prepare input messages with media uploads | ||
| const langfuseInput = await prepareLangfuseInput(args.inputMessages); | ||
|
|
||
| // Initialize Langfuse client and create trace | ||
| const langfuse = new Langfuse(); | ||
| const trace = langfuse.trace({ | ||
| name: "generation", | ||
| userId: args.userId ? String(args.userId) : undefined, | ||
| input: langfuseInput, | ||
| sessionId: args.sessionId, | ||
| }); | ||
|
|
||
| // Common trace metadata | ||
| const tags = [...(args.tags ?? []), ...extractTags(operationNode)]; | ||
| const metadata = { | ||
| ...args.metadata, | ||
| ...extractMetadata(operationNode), | ||
| }; | ||
|
|
||
| const llm = operationNode.content.llm; | ||
|
|
||
| const generationName = isTextGenerationNode(operationNode) | ||
| ? "generateText" | ||
| : isImageGenerationNode(operationNode) | ||
| ? "generateImage" | ||
| : undefined; | ||
|
|
||
| if (args.generation.status === "failed") { | ||
| trace.update({ | ||
| tags, | ||
| metadata, | ||
| }); | ||
|
|
||
| trace.generation({ | ||
| name: generationName, | ||
| model: llm.id, | ||
| modelParameters: llm.configurations, | ||
| input: langfuseInput, | ||
| startTime: new Date(args.generation.startedAt), | ||
| endTime: new Date(args.generation.failedAt), | ||
| metadata, | ||
| level: "ERROR", | ||
| statusMessage: args.generation.error.message, | ||
| }); | ||
| await langfuse.flushAsync(); | ||
| return; | ||
| } | ||
|
|
||
| // Handle text generation telemetry | ||
| if (isTextGenerationNode(operationNode)) { | ||
| const usage = args.generation.usage ?? { | ||
| inputTokens: 0, | ||
| outputTokens: 0, | ||
| totalTokens: 0, | ||
| }; | ||
|
|
||
| // Calculate costs for token-based models | ||
| const cost = await calculateDisplayCost(llm.provider, llm.id, { | ||
| inputTokens: usage.inputTokens ?? 0, | ||
| outputTokens: usage.outputTokens ?? 0, | ||
| }); | ||
|
|
||
| trace.update({ | ||
| output: args.generation.outputs, | ||
| tags, | ||
| metadata, | ||
| }); | ||
|
|
||
| trace.generation({ | ||
| name: generationName, | ||
| model: llm.id, | ||
| modelParameters: llm.configurations, | ||
| input: langfuseInput, | ||
| output: args.generation.outputs, | ||
| usage: { | ||
| unit: "TOKENS", | ||
| input: usage.inputTokens ?? 0, | ||
| output: usage.outputTokens ?? 0, | ||
| total: usage.totalTokens ?? 0, | ||
| inputCost: cost.inputCostForDisplay, | ||
| outputCost: cost.outputCostForDisplay, | ||
| totalCost: cost.totalCostForDisplay, | ||
| }, | ||
| startTime: new Date(args.generation.startedAt), | ||
| endTime: new Date(args.generation.completedAt), | ||
| metadata, | ||
| level: "DEFAULT", | ||
| }); | ||
| } | ||
|
|
||
| // Handle image generation telemetry | ||
| if (isImageGenerationNode(operationNode)) { | ||
| // Convert output files to Langfuse media references | ||
| const mediaReferences = (args.outputFileBlobs ?? []).map( | ||
| (file) => | ||
| new LangfuseMedia({ | ||
| contentType: file.contentType as ApiMediaContentType, | ||
| contentBytes: Buffer.from(file.bytes), | ||
| }), | ||
| ); | ||
|
|
||
| if (mediaReferences.length > 0) { | ||
| trace.update({ | ||
| output: mediaReferences, | ||
| tags, | ||
| metadata, | ||
| }); | ||
|
|
||
| trace.generation({ | ||
| name: generationName, | ||
| model: llm.id, | ||
| modelParameters: llm.configurations, | ||
| input: langfuseInput, | ||
| output: mediaReferences, | ||
| usage: { | ||
| input: 0, | ||
| output: 0, | ||
| total: 0, | ||
| unit: "IMAGES", | ||
| }, | ||
| startTime: new Date(args.generation.startedAt), | ||
| endTime: new Date(args.generation.completedAt), | ||
| metadata, | ||
| }); | ||
| } | ||
| } | ||
|
|
||
| await langfuse.flushAsync(); | ||
| } catch (error) { | ||
| // Log error with context for debugging | ||
| console.error("Telemetry emission failed:", { | ||
| error: error instanceof Error ? error.message : String(error), | ||
| nodeType: args.generation.context.operationNode.content.type, | ||
| generationId: args.generation.id, | ||
| }); | ||
| } | ||
| } | ||
| return trace; | ||
| }; |
There was a problem hiding this comment.
4. Telemetry emission removed 🐞 Bug ✧ Quality
packages/langfuse/src/trace-generation.ts now only creates and returns a trace and never records generation events or flushes, so prior generation telemetry (inputs/outputs/status/tags/cost/usage) is no longer emitted. The studio wrapper also only returns this trace without further updates, so nothing in the current code path emits generation telemetry.
Agent Prompt
## Issue description
`traceGeneration` no longer emits generation telemetry (no trace.update/generation/flush), and the studio wrapper doesn’t add those calls either, so generation telemetry is effectively disabled.
## Issue Context
Embedding telemetry still emits and flushes, suggesting generation telemetry should follow the same pattern.
## Fix Focus Areas
- packages/langfuse/src/trace-generation.ts[1-32]
- apps/studio.giselles.ai/lib/trace.ts[10-29]
- packages/langfuse/src/trace-embedding.ts[60-98]
ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools
🔍 QA Testing Assistant by Giselle📋 Manual QA ChecklistBased on the changes in this PR, here are the key areas to test manually:
✨ Prompt for AI AgentsUse the following prompts with Cursor or Claude Code to automate E2E testing: 📝 E2E Test Generation Prompt4. MCP Integration Guidelines
5. CI-Ready Code Requirements
|
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 3
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@apps/studio.giselles.ai/lib/trace.ts`:
- Around line 4-8: Validate the Langfuse configuration before constructing the
Langfuse instance: check process.env.LANGFUSE_SECRET_KEY,
process.env.LANGFUSE_PUBLIC_KEY, and process.env.LANGFUSE_BASE_URL and throw a
clear, actionable Error if any are missing (e.g., mention which specific env var
is absent and how to set it) so the Langfuse constructor (Langfuse) never
receives undefined values; update the code around the Langfuse instantiation to
perform this early validation and fail fast with descriptive messages.
- Around line 10-29: Several callsites still call the removed export
traceGenerationForTeam; replace those imports and calls to use the new trace
function and its parameter shape. Update any import that references
traceGenerationForTeam to import trace instead, and adjust call-sites to pass an
object with traceId, optional userId, sessionId, and metadata (instead of the
old team-based signature), then remove any now-unused team-related args; make
sure the invoked function is trace(...) (not traceGenerationForTeam) and that
you pass deploymentId via the environment as the trace implementation expects.
- Around line 1-8: This module initializes Langfuse with secret environment keys
(Langfuse, langfuse, traceGeneration); add the Next.js server-only guard by
importing/adding the "server-only" directive as the very first statement in the
file (before any other imports) so the module cannot be accidentally bundled for
client-side use; keep the rest of the file unchanged and ensure the directive
appears above the imports of Langfuse and traceGeneration.
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: ce89c6d9-e815-4236-aa37-764028d8d29a
📒 Files selected for processing (2)
apps/studio.giselles.ai/lib/trace.tspackages/langfuse/src/trace-generation.ts
Comment on lines
+1
to
+8
| import { Langfuse } from "langfuse"; | ||
| import { traceGeneration } from "@repo/langfuse/trace-generation"; | ||
|
|
||
| type TeamForPlan = Pick< | ||
| CurrentTeam, | ||
| "id" | "activeSubscriptionId" | "activeCustomerId" | "plan" | ||
| >; | ||
| const langfuse = new Langfuse({ | ||
| secretKey: process.env.LANGFUSE_SECRET_KEY, | ||
| publicKey: process.env.LANGFUSE_PUBLIC_KEY, | ||
| baseUrl: process.env.LANGFUSE_BASE_URL, | ||
| }); |
Contributor
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "== Direct imports of lib/trace from client modules =="
rg -n --type ts --type tsx 'from ["'\''].*lib/trace["'\'']|from ["'\''].*\/trace["'\'']' apps/studio.giselles.ai \
| cut -d: -f1 | sort -u \
| while read -r file; do
if rg -n '^\s*["'\'']use client["'\'']' "$file" >/dev/null; then
echo "CLIENT_IMPORT: $file"
fi
doneRepository: giselles-ai/giselle
Length of output: 283
🏁 Script executed:
#!/bin/bash
set -euo pipefail
# First, let's check the current content of the trace.ts file
echo "=== Current content of trace.ts ==="
cat -n apps/studio.giselles.ai/lib/trace.ts
echo ""
echo "=== Checking for direct imports of trace module ==="
# Fix: Use simpler regex pattern for ripgrep
rg 'from ["\x27].*trace["\x27]' apps/studio.giselles.ai --type ts --type tsx -n
echo ""
echo "=== Checking for imports with full path ==="
rg 'import.*from.*lib/trace' apps/studio.giselles.ai --type ts --type tsx -n
echo ""
echo "=== Checking for client components that might import trace ==="
rg '(use client|"use client")' apps/studio.giselles.ai --type ts --type tsx -l | head -20Repository: giselles-ai/giselle
Length of output: 987
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "=== Searching for imports of trace module ==="
# Search for imports of the trace module
rg 'from ["\x27].*trace["\x27]|import.*trace' apps/studio.giselles.ai -n | grep -i 'lib/trace\|/trace'
echo ""
echo "=== Checking if trace is imported in any client component ==="
# Find all client components first
find apps/studio.giselles.ai -type f \( -name "*.ts" -o -name "*.tsx" \) | while read -r file; do
if grep -q '"use client"' "$file" 2>/dev/null; then
# This is a client component - check if it imports trace
if grep -q 'trace' "$file" 2>/dev/null; then
echo "CLIENT_COMPONENT_WITH_TRACE: $file"
grep -n 'import.*trace\|from.*trace' "$file" 2>/dev/null || true
fi
fi
done
echo ""
echo "=== Searching all imports of trace anywhere ==="
rg 'import.*trace' apps/studio.giselles.ai -n | head -30Repository: giselles-ai/giselle
Length of output: 1219
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "=== Checking trigger/generate-content-job.ts ==="
head -30 apps/studio.giselles.ai/trigger/generate-content-job.ts
echo ""
echo "=== Checking app/giselle.ts ==="
head -50 apps/studio.giselles.ai/app/giselle.ts
echo ""
echo "=== Verifying if any of these importing files are client components ==="
for file in apps/studio.giselles.ai/trigger/generate-content-job.ts apps/studio.giselles.ai/app/giselle.ts; do
echo "File: $file"
grep '"use client"' "$file" 2>/dev/null && echo " -> HAS 'use client'" || echo " -> NO 'use client' (server-side)"
done
echo ""
echo "=== Checking what is exported from trace.ts ==="
grep '^export' apps/studio.giselles.ai/lib/trace.tsRepository: giselles-ai/giselle
Length of output: 3425
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "=== Full content of trace.ts ==="
wc -l apps/studio.giselles.ai/lib/trace.ts
echo ""
echo "=== Complete trace.ts file ==="
cat apps/studio.giselles.ai/lib/trace.ts
echo ""
echo "=== Searching for traceGenerationForTeam in the entire apps/studio.giselles.ai ==="
rg 'traceGenerationForTeam' apps/studio.giselles.ai -nRepository: giselles-ai/giselle
Length of output: 1541
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "=== Searching for traceGenerationForTeam export ==="
rg 'export.*traceGenerationForTeam' apps/studio.giselles.ai -n
echo ""
echo "=== Searching for traceGenerationForTeam definition anywhere in repo ==="
rg 'const traceGenerationForTeam|function traceGenerationForTeam|export.*traceGenerationForTeam' . -n | head -20
echo ""
echo "=== Check if there might be a re-export or aliasing in trace.ts ==="
rg 'traceGenerationForTeam' apps/studio.giselles.ai/lib/trace.ts -nRepository: giselles-ai/giselle
Length of output: 114
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "=== Checking if trace.ts might be exporting traceGenerationForTeam via star export or re-export ==="
cat apps/studio.giselles.ai/lib/trace.ts | grep -E 'export|traceGenerationForTeam' || echo "No match found"
echo ""
echo "=== Looking for any trace.ts files ==="
find apps/studio.giselles.ai -name "trace.ts" -o -name "trace.tsx"
echo ""
echo "=== Checking if there's an index file that might re-export ==="
ls -la apps/studio.giselles.ai/lib/ | grep -E 'index|trace'
echo ""
echo "=== Let's verify the actual import line from generate-content-job.ts ==="
sed -n '1,10p' apps/studio.giselles.ai/trigger/generate-content-job.tsRepository: giselles-ai/giselle
Length of output: 873
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "=== Searching for traceGenerationForTeam in entire repository ==="
rg 'traceGenerationForTeam' . -n 2>/dev/null | head -30
echo ""
echo "=== Checking `@repo/langfuse` package for this export ==="
find . -path "*/langfuse/*" -name "*.ts" -o -name "*.tsx" 2>/dev/null | head -20
echo ""
echo "=== Checking if there's a langfuse directory ==="
find . -type d -name "langfuse" 2>/dev/nullRepository: giselles-ai/giselle
Length of output: 2192
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "=== Looking for langfuse in packages directory ==="
find . -path "*/packages/*langfuse*" -type d 2>/dev/null | head -10
echo ""
echo "=== Checking monorepo structure ==="
ls -la packages/ 2>/dev/null | head -20
echo ""
echo "=== Search for where traceGenerationForTeam might be defined ==="
rg 'export.*function.*traceGenerationForTeam|export const traceGenerationForTeam' . 2>/dev/null
echo ""
echo "=== Check if perhaps the review comment is about a file that doesn't match current state ==="
echo "Looking for any recent changes or different trace files:"
find apps -name "*trace*" -type f 2>/dev/null | grep -v node_modulesRepository: giselles-ai/giselle
Length of output: 1402
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "=== Checking packages/langfuse structure ==="
find packages/langfuse -type f -name "*.ts" -o -name "*.tsx" 2>/dev/null
echo ""
echo "=== Checking main export from packages/langfuse ==="
cat packages/langfuse/src/index.ts 2>/dev/null || cat packages/langfuse/src/main.ts 2>/dev/null || ls -la packages/langfuse/src/
echo ""
echo "=== Search for traceGenerationForTeam in packages ==="
rg 'traceGenerationForTeam' packages/ 2>/dev/null
echo ""
echo "=== Check if trace.ts might be a recent addition and this is incomplete ==="
echo "Looking at the actual import in context:"
sed -n '1,25p' apps/studio.giselles.ai/app/giselle.ts | tail -10Repository: giselles-ai/giselle
Length of output: 433
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "=== Content of packages/langfuse/src/trace-generation.ts ==="
cat packages/langfuse/src/trace-generation.ts
echo ""
echo "=== Content of packages/langfuse/src/trace-embedding.ts ==="
head -50 packages/langfuse/src/trace-embedding.tsRepository: giselles-ai/giselle
Length of output: 1892
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "=== Comprehensive search for traceGenerationForTeam in entire repo ==="
rg 'traceGenerationForTeam' . --type ts --type tsx 2>/dev/null | head -30
echo ""
echo "=== Check if maybe it's defined in app/giselle.ts itself ==="
grep -n 'traceGenerationForTeam\|export.*trace' apps/studio.giselles.ai/app/giselle.ts | head -20
echo ""
echo "=== Check if there's aliasing happening in the imports ==="
grep -A5 'import.*trace' apps/studio.giselles.ai/app/giselle.ts | head -20
echo ""
echo "=== Double-check what's actually exported from trace.ts ==="
grep '^export' apps/studio.giselles.ai/lib/trace.tsRepository: giselles-ai/giselle
Length of output: 133
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "=== Checking the exact line where traceGenerationForTeam is imported ==="
sed -n '7p' apps/studio.giselles.ai/trigger/generate-content-job.ts
echo ""
echo "=== Checking if perhaps the file has been modified and export is missing ==="
echo "Full trace.ts again:"
cat -n apps/studio.giselles.ai/lib/trace.ts
echo ""
echo "=== Let's check if there are any TypeScript errors or if this even compiles ==="
echo "Checking tsconfig paths to understand module resolution:"
grep -r '"@/lib' apps/studio.giselles.ai/tsconfig.json 2>/dev/null || grep -r '"@/' tsconfig.json 2>/dev/null | head -5Repository: giselles-ai/giselle
Length of output: 1217
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "=== Checking git status to see if trace.ts is being modified ==="
cd apps/studio.giselles.ai && git diff lib/trace.ts 2>/dev/null || echo "No git diff available"
cd - > /dev/null
echo ""
echo "=== Let's verify what the actual issue is - check app/giselle.ts usage ==="
sed -n '228,235p' apps/studio.giselles.ai/app/giselle.ts
echo ""
echo "=== Let's see how traceGenerationForTeam is being called ==="
sed -n '228,245p' apps/studio.giselles.ai/app/giselle.tsRepository: giselles-ai/giselle
Length of output: 851
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "=== This is a critical issue - let's understand what traceGenerationForTeam should be ==="
echo "Based on usage patterns, checking if it's supposed to be a wrapper around trace:"
echo ""
echo "=== Check if there's a team parameter in the trace function ==="
sed -n '10,29p' apps/studio.giselles.ai/lib/trace.ts
echo ""
echo "=== Let's check what parameters traceGenerationForTeam receives ==="
sed -n '66,75p' apps/studio.giselles.ai/trigger/generate-content-job.ts
echo ""
echo "=== Summary: Is trace.ts exporting everything it should? ==="
echo "Expected exports from trace.ts based on imports:"
echo "1. trace (currently exported)"
echo "2. traceGenerationForTeam (NOT exported - MISSING!)"
echo ""
echo "=== This appears to be incomplete code - let's verify the review comment is about this file ==="
echo "Review comment references lines 1-8 of trace.ts, which matches what we see."Repository: giselles-ai/giselle
Length of output: 1399
Add "server-only" guard to prevent accidental client-side imports of secret environment variables.
This module initializes Langfuse with secret keys (LANGFUSE_SECRET_KEY, LANGFUSE_PUBLIC_KEY, LANGFUSE_BASE_URL). Add the "server-only" directive at the top to enforce server-side-only usage and align with Next.js best practices.
Proposed fix
+import "server-only";
import { Langfuse } from "langfuse";
import { traceGeneration } from "@repo/langfuse/trace-generation";📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| import { Langfuse } from "langfuse"; | |
| import { traceGeneration } from "@repo/langfuse/trace-generation"; | |
| type TeamForPlan = Pick< | |
| CurrentTeam, | |
| "id" | "activeSubscriptionId" | "activeCustomerId" | "plan" | |
| >; | |
| const langfuse = new Langfuse({ | |
| secretKey: process.env.LANGFUSE_SECRET_KEY, | |
| publicKey: process.env.LANGFUSE_PUBLIC_KEY, | |
| baseUrl: process.env.LANGFUSE_BASE_URL, | |
| }); | |
| import "server-only"; | |
| import { Langfuse } from "langfuse"; | |
| import { traceGeneration } from "@repo/langfuse/trace-generation"; | |
| const langfuse = new Langfuse({ | |
| secretKey: process.env.LANGFUSE_SECRET_KEY, | |
| publicKey: process.env.LANGFUSE_PUBLIC_KEY, | |
| baseUrl: process.env.LANGFUSE_BASE_URL, | |
| }); |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@apps/studio.giselles.ai/lib/trace.ts` around lines 1 - 8, This module
initializes Langfuse with secret environment keys (Langfuse, langfuse,
traceGeneration); add the Next.js server-only guard by importing/adding the
"server-only" directive as the very first statement in the file (before any
other imports) so the module cannot be accidentally bundled for client-side use;
keep the rest of the file unchanged and ensure the directive appears above the
imports of Langfuse and traceGeneration.
Comment on lines
+4
to
+8
| const langfuse = new Langfuse({ | ||
| secretKey: process.env.LANGFUSE_SECRET_KEY, | ||
| publicKey: process.env.LANGFUSE_PUBLIC_KEY, | ||
| baseUrl: process.env.LANGFUSE_BASE_URL, | ||
| }); |
Contributor
There was a problem hiding this comment.
Fail fast on missing Langfuse credentials with actionable errors.
Lines 5-6 currently pass env vars through without validation. If missing, telemetry failures become hard to diagnose.
Proposed fix
import { Langfuse } from "langfuse";
import { traceGeneration } from "@repo/langfuse/trace-generation";
+const getRequiredEnvVar = (name: string): string => {
+ const value = process.env[name];
+ if (!value) {
+ throw new Error(`Missing required environment variable: ${name}`);
+ }
+ return value;
+};
+
const langfuse = new Langfuse({
- secretKey: process.env.LANGFUSE_SECRET_KEY,
- publicKey: process.env.LANGFUSE_PUBLIC_KEY,
+ secretKey: getRequiredEnvVar("LANGFUSE_SECRET_KEY"),
+ publicKey: getRequiredEnvVar("LANGFUSE_PUBLIC_KEY"),
baseUrl: process.env.LANGFUSE_BASE_URL,
});🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@apps/studio.giselles.ai/lib/trace.ts` around lines 4 - 8, Validate the
Langfuse configuration before constructing the Langfuse instance: check
process.env.LANGFUSE_SECRET_KEY, process.env.LANGFUSE_PUBLIC_KEY, and
process.env.LANGFUSE_BASE_URL and throw a clear, actionable Error if any are
missing (e.g., mention which specific env var is absent and how to set it) so
the Langfuse constructor (Langfuse) never receives undefined values; update the
code around the Langfuse instantiation to perform this early validation and fail
fast with descriptive messages.
Comment on lines
+10
to
+29
| export const trace = ({ | ||
| traceId, | ||
| userId, | ||
| sessionId, | ||
| metadata, | ||
| }: { | ||
| traceId: string; | ||
| userId?: string; | ||
| sessionId?: string; | ||
| userId: string; | ||
| team: TeamForPlan; | ||
| providerMetadata?: ProviderMetadata; | ||
| requestId?: string; | ||
| }) { | ||
| const teamPlan = args.team.plan; | ||
| const planTag = `plan:${teamPlan}`; | ||
|
|
||
| await traceGeneration({ | ||
| generation: args.generation, | ||
| outputFileBlobs: args.outputFileBlobs, | ||
| inputMessages: args.inputMessages, | ||
| userId: args.userId, | ||
| tags: [planTag], | ||
| metadata: { | ||
| generationId: args.generation.id, | ||
| teamPlan, | ||
| userId: args.userId, | ||
| subscriptionId: args.team.activeSubscriptionId ?? "", | ||
| customerId: args.team.activeCustomerId ?? "", | ||
| providerMetadata: args.providerMetadata, | ||
| requestId: args.requestId, | ||
| workspaceId: args.generation.context.origin.workspaceId, | ||
| }, | ||
| sessionId: args.sessionId, | ||
| metadata?: Record<string, unknown>; | ||
| }) => { | ||
| return traceGeneration({ | ||
| langfuse, | ||
| traceId, | ||
| userId, | ||
| sessionId, | ||
| metadata, | ||
| deploymentId: process.env.VERCEL_DEPLOYMENT_ID, | ||
| }); | ||
| } | ||
| }; |
Contributor
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "== References to removed/old API =="
rg -n -C2 '\btraceGenerationForTeam\b' --type ts --type tsx
echo
echo "== Exports from trace helper modules =="
fd 'trace\.ts$' | while read -r file; do
echo "-- $file"
rg -n 'export\s+(const|function|async function)\s+(trace|traceGenerationForTeam)\b' "$file" || true
doneRepository: giselles-ai/giselle
Length of output: 143
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "== References to traceGenerationForTeam =="
rg -n -C2 'traceGenerationForTeam' --type ts || echo "No matches found"
echo
echo "== Checking trace.ts file =="
fd -e ts 'trace\.ts$' -x cat -n {}
echo
echo "== Checking generate-content-job.ts =="
fd -e ts 'generate-content-job\.ts$' -x cat -n {}Repository: giselles-ai/giselle
Length of output: 7926
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "== Looking for traceGenerationForTeam definition/export =="
rg -n 'export.*traceGenerationForTeam|function traceGenerationForTeam|const traceGenerationForTeam' --type ts -A 3
echo
echo "== Full contents of trace.ts =="
cat apps/studio.giselles.ai/lib/trace.ts
echo
echo "== Search in repo for traceGenerationForTeam definition =="
rg -n 'traceGenerationForTeam\s*[=:({]' --type ts | head -20Repository: giselles-ai/giselle
Length of output: 121
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "== Searching entire codebase for traceGenerationForTeam =="
rg 'traceGenerationForTeam' --type ts -l
echo
echo "== Check if trace.ts is the file under review =="
ls -la apps/studio.giselles.ai/lib/trace.tsRepository: giselles-ai/giselle
Length of output: 351
Critical breaking API change: 6 active callsites still reference removed traceGenerationForTeam export.
The trace() function replaces traceGenerationForTeam(), but 2 files still import and call the old API:
apps/studio.giselles.ai/trigger/generate-content-job.ts(lines 66, 86)apps/studio.giselles.ai/app/giselle.ts(lines 228, 244, 261, 276)
These callsites must be migrated to use the new trace() export to prevent compilation and runtime failures.
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@apps/studio.giselles.ai/lib/trace.ts` around lines 10 - 29, Several callsites
still call the removed export traceGenerationForTeam; replace those imports and
calls to use the new trace function and its parameter shape. Update any import
that references traceGenerationForTeam to import trace instead, and adjust
call-sites to pass an object with traceId, optional userId, sessionId, and
metadata (instead of the old team-based signature), then remove any now-unused
team-related args; make sure the invoked function is trace(...) (not
traceGenerationForTeam) and that you pass deploymentId via the environment as
the trace implementation expects.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fixes #2651
Added
deploymentIdfield to Langfuse trace metadata usingprocess.env.VERCEL_DEPLOYMENT_ID. This enables deployment-level tracking and debugging in Langfuse. Updated the trace generation function to accept the optional deploymentId parameter and pass it through from the trace wrapper.Summary by CodeRabbit