I have Telegraf running for telemetry streaming via GRPC, with the following telegraf.conf configuration:

`[agent]
  logtarget = "file"
  logfile = "/var/log/telegraf/telegraf-core.log"
  debug = true
  quiet = true
  round_interval = false
  metric_batch_size = 1000000
  metric_buffer_limit = 1000000

[[inputs.jti_openconfig_telemetry]]
  servers = ["x.x.x.x:50051","x.x.x.x:50051"]

  sample_frequency = "5000ms"
  username = "xxxx"
  password = "xxxx"
  client_id = "Telegraf"

  sensors = [
    "interfaces_oper_state /interfaces/interface/state/oper-status"
  ]`

The database I’m using is Prometheus, with the following configuration:

`[[outputs.prometheus_client]]
  listen = ":9200"`

This setup results in the following enumeration response:

interfaces_oper_state__component_id{device="x.x.x.x", host="grafana-poc", instance="x.x.x.x:9200", interfaces_interface__name="name", interfaces_interface_state_oper_status="UP", job="job", location="example", path="sensor_1011:/interfaces/interface/state/oper-status/:/interfaces/interface/state/oper-status/:mib2d", system_id="example"} 65535.

I want to create an alert based on the interfaces_interface_state_oper_status. To achieve this, I am using the [[processors.enum]] processor with the following configuration:

`[[processors.enum]]
  [[processors.enum.mapping]]
    field = "/interfaces/interface/state/oper-status"
    dest = "status_code"
    [processors.enum.mapping.value_mappings]
      "UP" = 0
      "DOWN" = 1
      "TESTING" = 2
      "UNKNOWN" = 3
      "DORMANT" = 4
      "NOT_PRESENT" = 5
      "LOWER_LAYER_DOWN" = 6`

This results in the following metric:

interfaces_oper_state_status_code{device="x.x.x.x", host="grafana-poc", instance="x:9200", interfaces_interface__name="name", interfaces_interface_state_oper_status="UP", job="Job", location="location", path="sensor_1011:/interfaces/interface/state/oper-status/:/interfaces/interface/state/oper-status/:mib2d", system_id="example"} 0.

This configuration works as expected; when I create a visualization in Grafana and shut down an interface, the value changes from 0 to 1.

However, when I try to create an alert based on this metric using the changes(interfaces_oper_state_status_code[$__interval]) function, I always get a value of 0 (I tested it with all interval values), even when I shut down or bring up an interface.
Although the value of interfaces_oper_state_status_code changes, the changes() function does not seem to work as intended.

I have tried other functions, like rate, delta, increase.

Since interfaces_oper_state_status_code changes in the right way when I bring up or shut down an interface, I am expecting that the result of changes(interfaces_oper_state_status_code[$__interval] would also work.