smart-agent_system-common: rename "per_cpu_enabled" variable to "agen…

…t_per_cpu_enabled" (#381)

modules/smart-agent_system-common/README.md

@@ -126,8 +126,8 @@ But we exclude it explicitly in related detector for safety to prevent any alert
 #### Load
 
 You have two choices to use load based detectors:
-  - either keep the `per_cpu_enabled` enabled (variable default value) __and__ define `perCPU: true` in the [load monitor](https://github.com/signalfx/signalfx-agent/blob/main/docs/monitors/load.md) configuration (for Kubernetes, you can use `loadPerCPU` option from the Helm chart available from `1.2.0` version).
-  - or override the `per_cpu_enabled` to `false` __and__ keep the default configuration for the [load monitor](https://github.com/signalfx/signalfx-agent/blob/main/docs/monitors/load.md) with `perCPU: false` or not defined
+  - either keep the `agent_per_cpu_enabled` enabled (variable default value) __and__ define `perCPU: true` in the [load monitor](https://github.com/signalfx/signalfx-agent/blob/main/docs/monitors/load.md) configuration (for Kubernetes, you can use `loadPerCPU` option from the Helm chart available from `1.2.0` version).
+  - or override the `agent_per_cpu_enabled` to `false` __and__ keep the default configuration for the [load monitor](https://github.com/signalfx/signalfx-agent/blob/main/docs/monitors/load.md) with `perCPU: false` or not defined
 
 In both cases, the goal is to get alerts based on the __ratio__ of load by dividing the original load per the number of CPU/cores which is the only way to get generic and relevant alerts for load.
 It mainly depends if you want to collect 2 metrics instead of 1 and if you want the load one to be raw or already averaged.

modules/smart-agent_system-common/conf/02-load.yaml

@@ -8,7 +8,7 @@ signals:
   num_processors:
     metric: cpu.num_processors
   signal:
-    formula: '(${var.per_cpu_enabled ? "load" : "load/num_processors"})'
+    formula: '(${var.agent_per_cpu_enabled ? "load" : "load/num_processors"})'
 rules:
   critical:
     threshold: 2.5

modules/smart-agent_system-common/conf/readme.yaml

@@ -31,8 +31,8 @@ source_doc: |
   #### Load
 
   You have two choices to use load based detectors:
-    - either keep the `per_cpu_enabled` enabled (variable default value) __and__ define `perCPU: true` in the [load monitor](https://github.com/signalfx/signalfx-agent/blob/main/docs/monitors/load.md) configuration (for Kubernetes, you can use `loadPerCPU` option from the Helm chart available from `1.2.0` version).
-    - or override the `per_cpu_enabled` to `false` __and__ keep the default configuration for the [load monitor](https://github.com/signalfx/signalfx-agent/blob/main/docs/monitors/load.md) with `perCPU: false` or not defined
+    - either keep the `agent_per_cpu_enabled` enabled (variable default value) __and__ define `perCPU: true` in the [load monitor](https://github.com/signalfx/signalfx-agent/blob/main/docs/monitors/load.md) configuration (for Kubernetes, you can use `loadPerCPU` option from the Helm chart available from `1.2.0` version).
+    - or override the `agent_per_cpu_enabled` to `false` __and__ keep the default configuration for the [load monitor](https://github.com/signalfx/signalfx-agent/blob/main/docs/monitors/load.md) with `perCPU: false` or not defined
 
   In both cases, the goal is to get alerts based on the __ratio__ of load by dividing the original load per the number of CPU/cores which is the only way to get generic and relevant alerts for load.
   It mainly depends if you want to collect 2 metrics instead of 1 and if you want the load one to be raw or already averaged.

modules/smart-agent_system-common/detectors-gen.tf

@@ -79,7 +79,7 @@ resource "signalfx_detector" "load" {
   program_text = <<-EOF
     load = data('load.midterm', filter=${module.filtering.signalflow})${var.load_aggregation_function}${var.load_transformation_function}
     num_processors = data('cpu.num_processors', filter=${module.filtering.signalflow})${var.load_aggregation_function}${var.load_transformation_function}
-    signal = (${var.per_cpu_enabled ? "load" : "load/num_processors"}).publish('signal')
+    signal = (${var.agent_per_cpu_enabled ? "load" : "load/num_processors"}).publish('signal')
     detect(when(signal > ${var.load_threshold_critical}, lasting=%{if var.load_lasting_duration_critical == null}None%{else}'${var.load_lasting_duration_critical}'%{endif}, at_least=${var.load_at_least_percentage_critical})).publish('CRIT')
     detect(when(signal > ${var.load_threshold_major}, lasting=%{if var.load_lasting_duration_major == null}None%{else}'${var.load_lasting_duration_major}'%{endif}, at_least=${var.load_at_least_percentage_major}) and (not when(signal > ${var.load_threshold_critical}, lasting=%{if var.load_lasting_duration_critical == null}None%{else}'${var.load_lasting_duration_critical}'%{endif}, at_least=${var.load_at_least_percentage_critical}))).publish('MAJOR')
 EOF

modules/smart-agent_system-common/variables.tf

@@ -1,7 +1,7 @@
 # Module specific
 
-variable "per_cpu_enabled" {
-  description = "Is perCPU option is enabled for the load monitor of the agent"
+variable "agent_per_cpu_enabled" {
+  description = "Is `perCPU` option is enabled for the load monitor in the agent configuration"
   type        = bool
   default     = true
 }