Laravel New Relic Laravel Package

Product Decisions This Supports

• Observability strategy / build vs. buy: This package is a thin, opinionated bridge between Laravel's queue and HTTP lifecycle and New Relic's PHP agent. Adopting it is a vote for "buy the APM vendor, keep the instrumentation glue thin and community-maintained" rather than building first-party transaction-naming and queue-tracing code in-house.
• Reliability and SLO reporting for background work: The v1.4.0 change moves queue transaction termination from JobProcessed to JobAttempted, so jobs that fail, throw, or are retried are closed out consistently instead of only on the happy path. This directly affects what your dashboards can claim about job success rate, retry volume, and queue latency — failed work shows up in New Relic as a distinct, attributable transaction rather than a hanging or misfiled one.
• Incident triage and on-call ergonomics: Queue jobs get a transaction name derived from the job class, and HTTP requests get named by route, so "which route/job is burning latency" is answerable without custom dashboards. This is the difference between an actionable alert and a wall of unknown transactions.
• Platform currency / upgrade roadmap input: The package now requires Laravel 11+. That's a concrete data point for the framework-upgrade roadmap — if you're still on Laravel 10, upgrading the APM integration is blocked until the app upgrade lands.
• Deployment and CI hygiene: Dependabot bumps (actions/checkout v6→v7, fetch-metadata 2.5→3.1) are maintenance-only, but they're a signal that the repo's supply chain is kept current — useful when arguing for the "safe to depend on" side of a dependency review.

When to Consider This Package

Consider adopting when:

• You are already a New Relic customer and want Laravel-aware transaction naming without writing custom middleware and queue event listeners yourself.
• Your application is on Laravel 11 or 12. v1.4.0 explicitly drops Laravel 10 (EOL), so the L10 upgrade is a hard prerequisite — check this before adding the dependency to composer.json, since it will otherwise fail resolution or pin you to v1.3.x.
• You run queue-heavy workloads and rely on per-job transaction data. The v1.4.0 fix is the reason to prefer this release over v1.3.0 if job failure visibility matters to you — staying on v1.3.0 means your failure-rate metrics are quietly under-reporting.
• You want the integration to be small and auditable — a handful of service providers and event hooks, not a framework of its own.

Look elsewhere when:

• You are on Laravel 10 and cannot upgrade in this cycle. Pin to v1.3.x and accept the JobProcessed-only transaction behavior, or budget the framework upgrade first.
• You use OpenTelemetry, Datadog, Sentry, or Grafana as your APM. There's no portable abstraction here; the package is New Relic-specific by design.
• You need deep custom instrumentation — span-level business metrics, per-tenant attribution, custom attributes on every job. You'll end up calling the New Relic agent API directly regardless, and the package only buys you the naming baseline.
• You want vendor-backed support and SLAs. This is a community package with a single primary maintainer; the v1.4.0 contributor (@alies-dev) is a first-timer, which is healthy but also a reminder of bus-factor risk.

How to Pitch It (Stakeholders)

To executives: A small open-source add-on that makes our New Relic APM actually understand our Laravel app — it names HTTP transactions by route and queue jobs by class, so when something slows down or breaks we can see which thing, not just that the app is unhealthy. The v1.4.0 release fixes a gap where failed background jobs weren't being reported properly, which means our job-reliability and failure-rate reporting becomes trustworthy rather than optimistic. Cost is negligible; the value is faster incident diagnosis and defensible SLO numbers.

To engineering: v1.4.0 is a small but meaningful correctness release. Queue transactions now end on JobAttempted instead of JobProcessed (PR #44, first-time contributor), which means failed and retried jobs close their transaction properly — a genuine fix for anyone tracking queue health. The trade-off: Laravel 10 support is dropped as EOL, so this is a breaking change at the dependency level. If we're on L10, we pin to v1.3.x or upgrade the framework; if we're on 11/12, bump and verify our queue dashboards now include failure data. Remaining changes are Dependabot maintenance on CI actions — low risk, but worth a glance at the workflow diff since actions/checkout v6→v7 can change default behavior.