Support custom recordingRules for Prometheus

Custom recordingRules can be added to the installation via
Helm for custom autoscaling rules.

See the README for more information.

Closes: #1225

Signed-off-by: Han Verstraete (OpenFaaS Ltd) <han@openfaas.com>
Signed-off-by: Alex Ellis (OpenFaaS Ltd) <alexellis2@gmail.com>

Author: welteki

chart/openfaas/Chart.yaml

@@ -2,7 +2,7 @@ apiVersion: v2
 type: application
 description: OpenFaaS - Serverless Functions Made Simple
 name: openfaas
-version: 14.2.72
+version: 14.2.73
 sources:
 - https://github.com/openfaas/faas
 - https://github.com/openfaas/faas-netes

chart/openfaas/README.md

@@ -409,7 +409,28 @@ OpenFaaS Pro will only scale down functions which have marked themselves as elig
 
 See also: [Scale to Zero docs](https://docs.openfaas.com/openfaas-pro/scale-to-zero/).
 
-## Removing the OpenFaaS
+## Custom autoscaling rules
+
+In order to build custom autoscaling rules, an additional recording rule is required for Prometheus for each type of scaling you want to add.
+
+To add latency-based scaling with the metrics recorded at the gateway, you could add the following to values.yaml:
+
+```yaml
+prometheus:
+  recordingRules:
+    - record: job:function_current_load:sum
+      expr: |
+        sum by (function_name) (rate(gateway_functions_seconds_sum{}[30s])) / sum by (function_name)  (rate( gateway_functions_seconds_count{}[30s]))
+        and on (function_name) avg by(function_name) (gateway_service_target_load{scaling_type="latency"}) > bool 1
+      labels:
+        scaling_type: latency
+```
+
+To check the configuration of current recording rules use the Prometheus UI or run `kubectl edit -n openfaas configmap/prometheus-config`.
+
+See also: [How to scale OpenFaaS Functions with Custom Metrics](https://www.openfaas.com/blog/custom-metrics-scaling/).
+
+## Removing OpenFaaS
 
 All control plane components can be cleaned up with helm:
 
@@ -671,3 +692,4 @@ For legacy scaling in OpenFaaS Community Edition.
 | `prometheus.retention.time` | When to remove old data from the prometheus db. | `15d` |
 | `prometheus.retention.size` | The maximum number of bytes of storage blocks to retain. Units supported: B, KB, MB, GB, TB, PB, EB. 0 meaning disabled. See: [Prometheus storage](https://prometheus.io/docs/prometheus/latest/storage/#operational-aspects)| `0` |
 | `prometheus.resources` | Resource limits and requests for prometheus containers | See [values.yaml](./values.yaml) |
+| `prometheus.recordingRules` | Custom recording rules for autoscaling. | `[]` |

chart/openfaas/templates/prometheus-pro-cfg.yaml

@@ -207,6 +207,10 @@ data:
         expr: ceil(sum(irate ( pod_cpu_usage_seconds_total{}[1m])*1000) by (function_name) * on (function_name) avg by (function_name) (gateway_service_target_load{scaling_type="cpu"}  > bool 1 ))
         labels:
           scaling_type: cpu
+      
+      {{- with .Values.prometheus.recordingRules }}
+      {{ toYaml . | nindent 6 }}
+      {{- end }}
 
     - name: recently_started_1m
       interval: 10s

chart/openfaas/values.yaml

@@ -360,6 +360,7 @@ prometheus:
       memory: "512Mi"
       cpu: "100m"
   annotations: {}
+  recordingRules: []
 
 alertmanager:
   image: prom/alertmanager:v0.27.0