concourse task wait metrics are missed

[dmytro-zhernosiekov-rft]

Summary

There are several metrics missing on our Concourse CI: concourse_tasks_waiting_duration concourse_tasks_waiting which really important for autoscalling and tracking user pipeline delays

Steps to reproduce

Deploy Concourse CI with Helm chart:

variable "concourse_version" {
  type        = string
  description = "Concourse release."
  default     = "6.7.5"
}

variable "helm_chart_version" {
  type    = string
  default = "11.4.1"
}

Expected results

concourse_tasks_watingin metrics appeared on prometheus service endpoint.

Actual results

Port forward:

kubectl port-forward service/concourse-web-prometheus -n concourse 9391:9391
Forwarding from 127.0.0.1:9391 -> 9391
Handling connection for 9391
Handling connection for 9391

In terminal:

$> curl http://localhost:9391/ | grep wait
# HELP go_memstats_heap_idle_bytes Number of heap bytes waiting to be used.
dmytro@centos: ~/Documents/refi/sdp-app-deploy $> 

Other metrics are present.

Additional context

Concourse run in AWS EKS, deployed with Helm Chart.

Triaging info

• Concourse version: 6.7.5
• Browser (if applicable): does not matter
• Did this used to work? no

[taylorsilva]

Duplicate of https://github.com/concourse/concourse/issues/6552 Keeping discussion in concourse/concourse since, if there is an issue, it's probably not in the helm chart.