Telegraf ships with a top-level [global_tags] block in its config format. Anything defined there gets stamped onto every metric the agent emits, no matter which input gathered it or which output it's shipped to. VyOS doesn't currently expose any way to set this — the telegraf subtree only lets you configure per-destination settings (influxdb, azure-data-explorer, splunk, prometheus-client, loki).

The gap matters for redundant deployments — active/standby pairs, clustered nodes, anything where the node doing the work can change over time. Right now the only identity a metric carries is host, so any time a role migrates to a different node, time-series continuity breaks and downstream dashboards/alerts built around a logical service identity (rather than a physical box) stop lining up.

suggested syntax:

set service monitoring telegraf global-tag <tag-name> value <tag-value>

This needs to live directly under telegraf, not under influxdb or any other output node — since global_tags in Telegraf applies before the metric ever reaches an output plugin, nesting it under one output would mean it silently stops working (or has to be duplicated) the moment a second output is added.

Relevant files:

interface-definitions/service_monitoring_telegraf.xml.in, data/templates/telegraf/telegraf.j2, src/conf_mode/service_monitoring_telegraf.py