Centrifugo Bundle Laravel Package

Product Decisions This Supports

• Real-time feature feasibility (build vs. buy vs. wire up): Whether live chat, in-app notifications, presence/typing indicators, live counters, or collaborative views are worth shipping now. Centrifugo is a standalone Go WebSocket server, so real-time fan-out no longer has to live inside the PHP request/response lifecycle — this unlocks scoping those features without standing up a Node/Socket.io service.
• Real-time transport sourcing: Self-hosted Centrifugo (with Redis/NATS as broker) vs. managed vendors (Pusher, Ably, Soketi, Laravel Reverb). This is a recurring-cost and vendor-lock-in decision, and it drives infra roadmap (ops ownership of the socket tier).
• Architecture direction: Decoupling the real-time layer from the monolith. PHP only publishes events; Centrifugo holds the connections. That means you can scale the app tier and the socket tier independently, and it constrains future "do we move to a separate realtime service?" debates.
• Stack-fit / dependency triage: This is a Symfony bundle, not a Laravel package — despite being filed here. That's a first-class decision input: it's only directly usable if the product is on Symfony. For a Laravel codebase, the correct decision is to call Centrifugo's HTTP/gRPC API directly (or use a Laravel broadcasting driver), not to adopt this.
• Adoption-risk posture: 11 stars, 0 dependents, single-vendor origin (Kismia's own bundle). This informs whether the team treats it as a dependency or as a reference implementation to copy/fork.

When to Consider This Package

Consider it only if:

• Your stack is Symfony, not Laravel. The value here is JWT auth-token generation plus a publish-API wrapper — roughly the surface area of a service class and a config block.
• You've already decided Centrifugo is the right real-time server, and you want the token/publish plumbing done idiomatically for Symfony's DI and config.
• You're comfortable pinning the version and accepting that you may need to fork it.

Look elsewhere (or skip the abstraction entirely) when:

• You're on Laravel. There's no service provider here. The integration is a handful of HTTP calls plus an HMAC/JWT signing step — cheaper to write and own than to wrap a foreign framework's bundle.
• You need maintenance guarantees. 0 dependents and a small star count means no de facto community support, no ecosystem pressure to keep it current with Centrifugo API changes (gRPC, presence, history, batching).
• You need more than basic publish + token. Advanced features (server-side subscriptions, channel history, presence inspection, gRPC transport) are likely thin or absent — go direct to Centrifugo's API.
• Your real-time needs are modest. If you're shipping notifications only, Laravel's broadcasting + Reverb/Pusher is a shorter path with far better documentation.

Practical recommendation: evaluate Centrifugo-the-server as the strategic decision; treat this bundle as disposable glue. Budget an afternoon to reimplement it rather than a dependency-review cycle to adopt it.

How to Pitch It (Stakeholders)

To executives: Real-time capability (chat, live notifications, presence) without per-connection SaaS pricing. Centrifugo is a mature, widely deployed Go server; we self-host it, pay for infrastructure rather than seats or concurrent-connection tiers, and keep the option to swap in a managed provider later. The specific PHP bundle in front of it is a small, replaceable convenience layer — it is not the strategic commitment, and it carries no meaningful vendor lock-in.

To engineering: The Centrifugo server is the decision. This bundle only wraps JWT token generation and the publish HTTP endpoint for Symfony — thin enough that forking or reimplementing it is low-cost, and low enough risk that we should treat it as a reference rather than a dependency. Note the framework mismatch: on Laravel we'd write the ~100-line client ourselves or use a broadcasting driver. Two constraints to accept going in: pin the version, and assume we own maintenance.

Bottom line: Buy the server, not the bundle. Score 0.085 / 11 stars / 0 dependents makes this a poor candidate for a dependency-review gate — evaluate it as sample code, decide on Centrifugo on its own merits.