feat(refinery): Set ShutdownDelay based on terminationGracePeriodSeco…

…nds (#382)

## Which problem is this PR solving?

Refinery 2.8 attempts to drain data to remaining peers on shutdown, and
has a configurable ShutdownDelay field where it tries to process
remaining work while shutting down. In k8s we need this config option to
be inline with terminationGracePeriodSeconds to ensure k8s doesn't kill
pods early.

it's the same change as #360 
## Short description of the changes
- added terminationGracePeriodSeconds configuration option in
values.yaml
- set ShutdownDelay to be terminationGracePeriodSeconds - 5 seconds by
default. Users can override this value in their values.yaml
- publish a `-hnyinternal` tag for the change


## How to verify that this has the expected result
Test localling using kind

charts/refinery/Chart.yaml

@@ -2,8 +2,8 @@ apiVersion: v2
 name: refinery
 description: Chart to deploy Honeycomb Refinery
 type: application
-version: 2.11.1
-appVersion: 2.7.1
+version: 2.11.2-hnyinternal.0
+appVersion: 2.8.0-dev.0
 keywords:
   - refinery
   - honeycomb

charts/refinery/templates/deployment.yaml

@@ -49,6 +49,7 @@ spec:
         {{- toYaml . | nindent 8 }}
       {{- end }}
     spec:
+      terminationGracePeriodSeconds: {{ .Values.terminationGracePeriodSeconds }}
       {{- with .Values.imagePullSecrets }}
       imagePullSecrets:
       {{- toYaml . | nindent 8 }}

charts/refinery/values.yaml

@@ -24,6 +24,7 @@ config:
     Host: '{{include "refinery.redis.fullname" .}}:6379'
 
   Collection:
+    ShutdownDelay: '{{ sub .Values.terminationGracePeriodSeconds 5 }}s'
     # AvailableMemory is the amount of system memory available to the Refinery process.
     AvailableMemory: '{{ .Values.resources.limits.memory }}'
     MaxMemoryPercentage: 75
@@ -319,6 +320,8 @@ tolerations: []
 
 affinity: {}
 
+terminationGracePeriodSeconds: 35
+
 # PodDisruptionBudget:
 #
 # Limit the number of concurrent disruptions that your application experiences, allowing for higher availability while