Collector Laravel Package

Product Decisions This Supports

• Build vs. buy for runtime observability. The package ships ready-made collectors for Redis, FrankenPHP, and Caddy. Adopting it replaces bespoke health-check and version-scraping code you would otherwise own, test, and maintain across every Laravel service.
• Platform roadmap bets. If FrankenPHP/Caddy is on your migration path (or already in production), this gives you instrumentation on day one rather than after the migration. The v2.3.1 fix for vv-prefixed version strings matters specifically for PaaS builds (e.g. Clever Cloud), so it de-risks "we can't see versions on our host" findings during a platform review.
• Managed-service compatibility. The Redis fix — a failing CONFIG GET no longer crashes collection when no logger is supplied — is a direct signal that the package is being hardened for managed Redis (ElastiCache, Upstash, Heroku, Scalingo) where CONFIG is commonly disabled. That turns "our managed Redis breaks the dashboard" from a blocker into a non-event.
• Operational readiness and incident response. These collectors feed dashboards and alerting. The new behavior of logging unparseable version output (truncated to 500 chars) means a missing version becomes a diagnosable gap instead of a silent hole in telemetry — supportable by on-call without a code change.
• SLO/health-dashboard scope. Concrete use cases: version-drift tracking across FrankenPHP/Caddy fleets, Redis health panels, and pre/post-deploy verification on nonstandard hosting.
• Support-cost and upgrade decisions. This is a patch release with no breaking changes and no configuration or migration steps. It is a low-risk, high-signal upgrade; there is no reason to schedule it behind a feature freeze.

When to Consider This Package

Consider adoption when:

• You run Laravel on FrankenPHP and/or Caddy and want version and runtime visibility without writing your own collectors.
• Your Redis is managed or locked down, so CONFIG GET may be unavailable — v2.3.1 explicitly handles this without crashing.
• You deploy to PaaS or container platforms with nonstandard build/version strings (doubled prefixes, custom output formats) and have been getting blank version fields.
• You want telemetry that degrades gracefully and leaves an audit trail in logs rather than failing collection silently.
• You already run a metrics/dashboard stack this package plugs into, so collectors are additive rather than a new dependency to operate.

Look elsewhere when:

• You need full APM, distributed tracing, or span-level profiling — reach for OpenTelemetry, an APM vendor, or Laravel Telescope/Octane tooling rather than collectors.
• You cannot host the storage/monitoring backend the package's data lands in; the collectors are only as useful as the store behind them.
• You need enterprise SLAs, long-term support guarantees, or a large maintainer team — this is a small, actively patched package and should be treated as such in your dependency risk register.
• Your stack is outside the supported runtimes (nginx/php-fpm only, unmanaged Redis) — the value of these specific collectors drops sharply.
• You require strict semantic-version pinning with a change-advisory process; the pace of small patch releases means you should automate dependency updates rather than hand-review each one.

How to Pitch It (Stakeholders)

For executives: This is a low-cost reliability improvement with no migration effort. The latest patch hardens the package against real infrastructure quirks — locked-down managed Redis and PaaS builds that report versions in nonstandard formats — so our monitoring stops going blind exactly where we have the least control. No breaking changes, no config changes; it is a routine upgrade that reduces the chance of a monitoring gap during an incident.

For engineering: Adopt the collectors instead of maintaining custom version-scraping and Redis health checks. v2.3.1 makes three things true that were previously fragile: a disabled CONFIG GET no longer throws during collection, Clever Cloud-style vv1.12.7 version strings now parse correctly, and unparseable version output is logged (truncated to 500 chars) so a missing version is visible and debuggable rather than silent. Upgrade is patch-level and safe to fold into the next deploy; the only behavioral change is additional notice-level logging on failure paths.