What problem does it solve? Adding monitoring, distributed tracing, and failure alerting to funboost message queue consumers normally requires modifying broker internals. This Skill shows how to attach production observability to any @boost-decorated function declaratively through Mixin classes and BoosterParams fields, without touching framework source code. ## Core Features & Use Cases - Prometheus Metrics: Record task counts, latency histograms, retries, and queue backlog via PrometheusConsumerMixin, with HTTP server mode for single-process and Pushgateway mode for multi-process consumers. - OpenTelemetry Tracing: Propagate W3C trace context across publishers and consumers with AutoOtelConsumerMixin and AutoOtelPublisherMixin, exporting spans to Jaeger or any OTLP-compatible backend. - Failure Alerting & Quotas: Send DingTalk, WeChat, Feishu, or webhook alerts on consecutive failures or error-rate thresholds with AlertNotifierConsumerMixin, and cap executions per period with PeriodicQuotaConsumerMixin. - Result Persistence: Save task status and return values to MongoDB via FunctionResultStatusPersistanceConfig, queryable in the funweb admin UI and usable with MongoAlertMonitor for aggregated alerts. - Use Case: An order-processing queue needs Grafana dashboards, cross-service tracing, and a WeChat alert after 5 consecutive failures. Combine the Prometheus, OTel, and AlertNotifier mixins through multiple inheritance and declare everything in one BoosterParams. ## Quick Start Add Prometheus metrics and failure alerting to my funboost task by configuring the appropriate consumer and publisher mixins in its BoosterParams.
