启用 Prometheus 指标#

为收集并暴露指标数据，n8n 采用 prom-client 库实现相关功能。

默认情况下，/metrics 端点处于禁用状态，但可通过 N8N_METRICS 环境变量启用该功能。

export N8N_METRICS=true

请参考相应的环境变量（N8N_METRICS_INCLUDE_*）来配置需要暴露的指标和标签。

主实例和工作实例均可暴露指标。

队列指标#

要启用队列指标，需将环境变量 N8N_METRICS_INCLUDE_QUEUE_METRICS 设置为 true。可通过 N8N_METRICS_QUEUE_METRICS_INTERVAL 调整刷新频率。

n8n 从 Bull 收集这些指标并在主实例上暴露。在多主实例配置中，进行查询聚合时，可通过值为 1 的 instance_role_leader 计量器识别主节点（非主节点该值为 0）。

# HELP n8n_scaling_mode_queue_jobs_active Current number of jobs being processed across all workers in scaling mode.
# TYPE n8n_scaling_mode_queue_jobs_active gauge
n8n_scaling_mode_queue_jobs_active 0

# HELP n8n_scaling_mode_queue_jobs_completed Total number of jobs completed across all workers in scaling mode since instance start.
# TYPE n8n_scaling_mode_queue_jobs_completed counter
n8n_scaling_mode_queue_jobs_completed 0

# HELP n8n_scaling_mode_queue_jobs_failed Total number of jobs failed across all workers in scaling mode since instance start.
# TYPE n8n_scaling_mode_queue_jobs_failed counter
n8n_scaling_mode_queue_jobs_failed 0

# HELP n8n_scaling_mode_queue_jobs_waiting Current number of enqueued jobs waiting for pickup in scaling mode.
# TYPE n8n_scaling_mode_queue_jobs_waiting gauge
n8n_scaling_mode_queue_jobs_waiting 0