gRPC and tRPC call a function on another machine. Moleculer is the service layer around that call. Same service 3 ways, same 4 problems: a 2nd instance, a dead one, a slow one, a new one. Plus how to combine them. dev.to/icebob/moleculer-vs-g… #nodejs #grpc

Events are how a microservice system stops being a distributed monolith. Emit vs broadcast, consumer groups, the same Moleculer code on NATS, Redis and Kafka, plus durable channels with retry + dead-letter. dev.to/icebob/event-driven-a… #nodejs #kafka

A dependency that is almost working is worse than one that is down. In Moleculer, retry, circuit breaker and timeout are broker options, not three more libraries. 6 live demos, incl. a breaker that cuts off one bad node, then heals. dev.to/icebob/circuit-breake… #nodejs

Where is the service, and which copy do I call? In Moleculer every node carries its own registry: no Consul, no etcd, no Kubernetes. 5 live demos: scaling, crashes under traffic, balancing strategies, discovery over plain TCP. dev.to/icebob/service-discov… #nodejs #microservices

🔒 Security release: Moleculer 0.15.2 & 0.14.36 fix an unauthenticated DoS vulnerability in the TCP transporter packet parser (GHSA-h89v-g948-47vp, CVSS 7.5 High). If you use the TCP transporter, please upgrade ASAP. Advisory: github.com/moleculerjs/molec…

Moleculer v0.15.1 is out. Key fix: zombie nodes no longer cause 404s in moleculer-web's autoAliases — the service registry now regenerates on disconnect instead of holding stale routes. Plus cluster signal forwarding + Node 22 crypto fixes.

In Moleculer, broker.call("users.get", { id: 1 }) looks the same whether "users" runs in-process or on another machine. The framework handles routing, load balancing, and network calls transparently. Your service code never changes. #nodejs #microservices

Moleculer services can declare their dependencies and wait for them. dependencies: ["users", "payments"] Your service will not start until they are available. No race conditions, no manual health polling, no startup ordering scripts. #nodejs #microservices

Building AI features into your backend? Moleculer's action-based architecture makes it simple. Wrap your LLM calls in a service action. Get retries, timeouts, circuit breaking, and load balancing for free. AI is just another action. #nodejs #AI

Moleculer action hooks let you run logic before and after any action. Auth check, input transform, audit log. No middleware chain, no decorators. hooks: { before: { "*": authCheck }, after: { "*": auditLog } } Clean, composable, built in. #nodejs #microservices

If you could add ONE feature to your microservice framework, what would it be? We are planning the next Moleculer release and genuinely want to hear what is missing from your stack.

Run v1 and v2 of the same service side by side in one Moleculer cluster. broker.call("v1.users.list") broker.call("v2.users.list") Gradual migration, zero downtime. Old clients keep working while new ones upgrade at their pace. #nodejs #microservices

Did you know? Moleculer handles graceful shutdown automatically. On SIGTERM it stops accepting new requests, waits for running actions to finish, disconnects from the transporter, then exits cleanly. Zero dropped requests during deploys. #nodejs #microservices

Did you know? In Moleculer, ctx.meta flows through the entire call chain automatically. Set auth info, tenant ID, or locale once — every downstream action inherits it. No need to pass it manually through 5 layers of calls. #nodejs #microservices

Every Moleculer action has built-in parameter validation. No middleware, no decorators. params: { email: "email", age: "number|min:18" } Invalid input? Rejected before your handler even runs. Supports nested objects, arrays, enums, custom rules. #nodejs #microservices

Need persistent pub/sub for your Moleculer services? @moleculer/channels gives you reliable messaging with Redis Streams, NATS JetStream, Kafka, or AMQP. Messages survive restarts, get retried on failure, and scale across consumer groups. npm i @moleculer/channels

Did you know? Moleculer's $node.health action gives you live CPU, memory, OS, and event-loop stats from any node — no extra middleware needed. Built-in observability, free. #nodejs #microservices

Moleculer has 4 built-in load balancing strategies for service instances: → Round Robin → Random → CPU-usage aware → Latency-based No external load balancer needed. Just run multiple instances and Moleculer handles the rest. #nodejs #microservices

Moleculer has built-in distributed caching with zero config: - Memory, Redis, or custom store - Per-action TTL and key generation - Auto-invalidation via event-based cache cleaning - Works across all nodes automatically One line in your action schema. #nodejs #microservices

Moleculer's built-in fault tolerance requires zero extra libraries: ⚡ Circuit Breaker 🔄 Retry with backoff ⏱️ Per-action timeouts 🪣 Bulkhead limiting Just add it to your broker config and you're protected. #nodejs #microservices