[ LISTICLE ]
Vapor runs on AWS Lambda. Persistent-agent APMs don't fit the execution model. Here's what actually works.
QUICK ANSWER
What's the best APM for Laravel Vapor?
Laravel Nightwatch Cloud is the officially-supported pick — same Laravel team, designed with Vapor in mind, zero config. Sentry is a strong alternative for polyglot stacks. AppSignal and Datadog (with Lambda Extension) also work. CloudWatch Insights covers basic log search. NightOwl's long-running agent architecture doesn't fit Lambda today — it's not in this list for that reason. Pick by whether you need cross-service tracing, source-map JS errors, or just Laravel-native APM.
Updated · 2026-04-13
| # | Tool | Fit with Vapor | Price |
|---|---|---|---|
| 1 | Laravel Nightwatch Cloud | SDK runs inside Lambda without a separate agent process. Official integration path. | $0 (300K events/mo free) → usage-based |
| 2 | Sentry | Works cleanly inside Lambda. Trace context propagates across AWS services if configured. | $26/mo → usage-based (per event + per span) |
| 3 | AppSignal | SDK works in Lambda. Cold-start tracking requires some manual instrumentation. | $25/mo → usage-based |
| 4 | Datadog APM + Lambda Extension | Lambda Extension runs alongside your function, handling the agent-lifecycle gap. Adds ~100ms cold-start overhead. | $31/host/mo + Lambda-specific line items |
| 5 | CloudWatch + Insights | Default destination for Lambda stdout/stderr. Zero configuration. | $0.50/GB ingested + query cost |
#1
Official first-party APM — designed with Vapor as a first-class deployment target.
PRICE
$0 (300K events/mo free) → usage-based
VAPOR FIT
SDK runs inside Lambda without a separate agent process. Official integration path.
BEST FOR
Vapor users who want zero-config, officially-supported Laravel APM.
#2
HTTP-based SDK, Lambda-safe, polyglot. Strong error tracking + performance spans.
PRICE
$26/mo → usage-based (per event + per span)
VAPOR FIT
Works cleanly inside Lambda. Trace context propagates across AWS services if configured.
BEST FOR
Vapor apps that need cross-service tracing into other AWS services.
#3
HTTP-based multi-language APM, Lambda-compatible PHP SDK.
PRICE
$25/mo → usage-based
VAPOR FIT
SDK works in Lambda. Cold-start tracking requires some manual instrumentation.
BEST FOR
Teams wanting an alternative to Sentry with a friendlier pricing curve at moderate scale.
#4
Enterprise APM with a dedicated Lambda Extension solving the agent-process problem.
PRICE
$31/host/mo + Lambda-specific line items
VAPOR FIT
Lambda Extension runs alongside your function, handling the agent-lifecycle gap. Adds ~100ms cold-start overhead.
BEST FOR
Enterprises already on Datadog for infra and wanting APM in the same pane.
#5
AWS's native logs + querying. Free-ish until you actually use it.
PRICE
$0.50/GB ingested + query cost
VAPOR FIT
Default destination for Lambda stdout/stderr. Zero configuration.
BEST FOR
Teams who only need log search and are fine without structured APM aggregation.
NOTE
NightOwl's agent is a long-running ReactPHP process that buffers telemetry and streams it to PostgreSQL. That design is fast (13,400 payloads/s single-instance) and cheap at scale — but it assumes a persistent worker, which Lambda doesn't provide. Lambda-compatible ingestion is something we're evaluating but haven't shipped.
For Vapor, pick from the options above. If you want BYOD PostgreSQL and Lambda becomes a hard requirement, let us know — it informs our roadmap.
NightOwl's architecture assumes a persistent agent process — a ReactPHP long-running worker that buffers telemetry and streams it out of band. That model doesn't fit AWS Lambda, which is Vapor's underlying execution environment. Lambda-compatible ingestion is something we're evaluating, but not shipping today. If you're committed to Vapor, pick from the options above.
Laravel Nightwatch Cloud is the officially-supported, zero-config pick from the same team. Sentry is a strong second for polyglot stacks or teams that want source-map JS error tracking. AppSignal is a credible middle-ground. Datadog if you're already on Datadog for infra.
Detect cold starts in a middleware (the first request a Lambda container handles is a cold start) and tag the telemetry event. Split p95 by cold vs warm — they tell different stories. See our Vapor monitoring guide for the code.
Vapor's dashboard shows Lambda execution metrics (invocation count, duration, errors per function) via CloudWatch integration. That's useful but not a full APM — no query pattern grouping, no grouped exceptions with alerts, no per-route p95 aggregation. Layer a real APM on top.
Limited options. You can ship CloudWatch logs to a self-hosted Loki or OpenSearch cluster for log aggregation. For traces, OpenTelemetry PHP SDK works in Lambda and can ship to a self-hosted Tempo. But operating a production observability stack alongside Vapor's serverless model defeats some of Vapor's value. Most Vapor users stay with SaaS APM.
Per-event pricing can compound quickly. A Vapor app doing 10M requests/month with full instrumentation typically lands at $200-500/mo for APM alone, on top of Lambda costs. Sampling helps but adds complexity. This is one area where BYOD models like NightOwl would be attractive if supported — but aren't yet.
PRICING
14-day free trial, no credit card. Your PostgreSQL, your data.
HOBBY
1 app · 14 days lookback · all Laravel events
TEAM
Up to 3 connected apps · unlimited environments · all Laravel events
AGENCY
Unlimited apps · unlimited agent instances · same flat rate at any traffic
Same package, your database
Laravel-native vs polyglot APM
$15 flat vs $99 per host
Full APM vs errors-only
Full APM vs exception-only
Production monitor vs local debugger
BYOD flat vs cloud tiered
BYOD flat vs cloud per-event
Full APM vs queue-only alerts
Laravel-focused vs enterprise platform
$15 flat vs per-host + per-product
Laravel-native vs polyglot APM
Full APM vs errors + uptime + cron
Full APM vs error tracking only
Laravel APM vs logs + uptime
Opinionated APM vs event platform
Structured APM vs log aggregator
Self-hosted options ranked
Full landscape in 2026