AI SDK Telemetry

Live analytics from Upstash Redis Search aggregations — last 24 hours.

Telemetry in code

Add the integration to your AI SDK calls, then read it back — every chart here is one Redis Search aggregation.

import { generateText } from "ai";
import { openai } from "@ai-sdk/openai";
import { redisSearchTelemetry } from "@/src/telemetry";

await generateText({
  model: openai("gpt-4o-mini"),
  prompt,
  experimental_telemetry: {
    isEnabled: true,
    functionId: "weather-bot",
    integrations: [redisSearchTelemetry()],
  },
});

Index

Active

nameai-telemetry

typejson

fields11

generations69

tool calls50

Generate telemetry

Run the AI SDK with the integration wired in — events land in the index and the charts update.

Runs 5 sample prompts.

Set OPENAI_API_KEY in the backend to enable data ingestion.

Generations

42 in last 30 min

Total tokens

52,983

35,641 in · 17,342 out

Tool calls

6 failed

Avg tokens / gen

768

across all generations

Tool latency

p50 / p95 / p99 per tool (ms), from $percentiles.

Tokens per agent

Average tokens from a $stats aggregation grouped by agent (functionId).

Finish reasons

Generation outcomes grouped with $terms.

Recent generations

Ordered by ts DESC — no sorted sets.

Agent	Model	Tokens	Finish	Time
chat-handler	gpt-4o	774	stop	12:39:03 PM
weather-bot	gpt-4o-mini	1,152	tool-calls	12:29:03 PM
summarizer	gpt-4o-mini	415	stop	12:27:03 PM
essayist	gpt-4o	1,297	stop	12:26:03 PM
weather-bot	gpt-4o-mini	589	stop	12:24:03 PM
weather-bot	gpt-4o-mini	1,099	stop	12:24:03 PM
status-bot	gpt-4o-mini	512	stop	12:15:03 PM
summarizer	gpt-4o-mini	370	stop	12:13:03 PM
status-bot	gpt-4o-mini	636	stop	12:02:03 PM
status-bot	gpt-4o-mini	707	stop	11:55:03 AM