Serverless observability
Debug route handlers, scheduled runs, and async jobs in one place — logs, traces, execution history, and latency views shared across every serverless entry point.
Related guides and docs
Get started free →How it works
Zero-config, full visibility
Invoke
Run your function from the UI, CLI, API gateway, or a pipeline — every invocation is captured automatically.
Trace
Each invocation produces a detailed trace with timing, logs, output, and the error stack when something fails.
Analyze
Aggregate metrics over time, and let the emerging insight helpers point at anomalies and slow functions — treat their suggestions as hints and validate them against your own baselines.
Features
Trace every invocation
Execution traces
Every invocation gets a full trace — steps, log lines, timing, input and output, and error details.
Live streaming
Run status streams live over SSE as your function executes — no polling, no page refreshes.
Latency percentiles
p50 / p95 / p99 latency updates after every run, so you spot regressions before your users do.
Run insights
Heuristics over your run history suggest where latency clusters — pair them with your own reviews before treating the output as authoritative.
Example
Execution trace payload
{ "runId": "run_01hw3k...", "functionName": "ai-summarizer", "status": "SUCCEEDED", "durationMs": 312, "startedAt": "2025-03-23T10:41:22.100Z", "logs": [ { "level": "INFO", "message": "Fetching URL..." }, { "level": "INFO", "message": "Calling OpenAI API..." }, { "level": "INFO", "message": "Summary generated (824 words)" } ], "output": { "summary": "AI funding surges as...", "wordCount": 824 } }