Langfuse insights

Open SourceVersions: []38 metrics

OpenTelemetry·

Dynatrace·Native·

Automated rollback prevents prompt deployment failures from impacting productioncritical

langfuse.api.error.rate langfuse.generation.latency langfuse.score.value

knowledge.businesscompassllc.com

14d ago▸

Response latency degradation indicates prompt performance issueswarning

langfuse.api.request.duration langfuse.generation.latency langfuse.api.error.rate

knowledge.businesscompassllc.com

14d ago▸

Token usage approaching budget limits triggers cost overrun riskwarning

langfuse.api.usage.tokens.total langfuse.api.usage.cost.total langfuse.generation.tokens.completion+1 more

knowledge.businesscompassllc.com

14d ago▸

Quality degradation detected through automated monitoring thresholdswarning

langfuse.score.value langfuse.api.error.rate

knowledge.businesscompassllc.com

14d ago▸

Test suite failures prevent unsafe prompt deploymentswarning

knowledge.businesscompassllc.com

14d ago▸

Edge case failures in production expose untested input variationswarning

knowledge.businesscompassllc.com

14d ago▸

Missing orderBy parameter causes 20x slowdown in traces.metrics endpoint

langfuse.api.request.duration langfuse.database.query.duration

19d ago▸

Synchronous tracing increases P99 latency from 200ms to 800mswarning

langfuse.trace.duration langfuse.api.request.duration

22d ago▸

Untracked prompt changes cause hallucination rate surge from 2% to 15%critical

langfuse.score.value

22d ago▸

LLM monthly cost exceeds 90% of budget

langfuse.api.usage.cost.total langfuse.api.usage.tokens.total langfuse.generation.tokens.prompt+1 more

22d ago▸

Self-hosted LangFuse requires manual scaling for high trafficinfo

langfuse.ingestion.processing.duration langfuse.ingestion.rate langfuse.database.connections.active

22d ago▸

Automated prompt deployment requires quality gate evaluationinfo

langfuse.score.value

22d ago▸

Readiness probe fails due to missing /api/public/ready endpointwarning

24d ago▸

UI lag when switching between observations in dashboardwarning

1mo ago▸

V1 observations endpoint experiences 10x latency increase after v2 release

langfuse.api.request.duration

3mo ago▸

Cloudflare proxy failure causes complete API and UI unavailability

langfuse.api.error.rate langfuse.http.request.count

4mo ago▸

Synthetic API tests detect health endpoint failurescritical

langfuse.api.request.count

4mo ago▸

Ingestion failures cause permanent trace data losscritical

langfuse.ingestion.events.total langfuse.traces.ingested langfuse.api.error.rate

4mo ago▸

SDK caching enables prompt access during API unavailability but fails on restartwarning

4mo ago▸

Cloudflare dashboard unavailability prevents DNS and proxy configuration changes

4mo ago▸

Status page provider dependency on same infrastructure delays incident communication

4mo ago▸

DNS-based bypass to AWS ALBs provides emergency mitigation pathinfo

4mo ago▸

Dashboard query timeout on large observations table prevents trace latency panel loading

langfuse.database.query.duration langfuse.observation.count

5mo ago▸

LEFT JOIN in trace percentile queries causes performance degradation at scale

langfuse.database.query.duration

5mo ago▸

ClickHouse queries without raw timestamp filters bypass index optimization

5mo ago▸

Worker container high memory usage with low CPU under ingestion load

langfuse.container.memory.utilization langfuse.container.cpu.utilization langfuse.worker.queue.depth

6mo ago▸

Redis OOM blocks BullMQ job processing and ingestion

langfuse.worker.queue.depth

6mo ago▸

Uneven memory distribution across worker replicaswarning

langfuse.container.memory.utilization

6mo ago▸

S3 rate limits cause worker job failures and dead letter queue accumulation

6mo ago▸

Web container high memory usage with low CPU utilizationwarning

langfuse.container.memory.utilization langfuse.container.cpu.utilization

6mo ago▸