chore: support to deploy ingress-apisix through helm charts #153
Conversation
Codecov Report
@@ Coverage Diff @@
## master #153 +/- ##
===========================================
+ Coverage 36.27% 50.35% +14.07%
===========================================
Files 28 30 +2
Lines 1257 1414 +157
===========================================
+ Hits 456 712 +256
+ Misses 771 618 -153
- Partials 30 84 +54
Continue to review full report at Codecov.
|
charts/base/templates/rbac.yaml
Outdated
| apiVersion: v1 | ||
| kind: ServiceAccount | ||
| metadata: | ||
| name: {{ .Values.rbac.serviceAccount }} |
There was a problem hiding this comment.
this name could be fixed format, such as {{ .Release.Name }} or {{ .Release.Name }}-svcacc, it will make values.yaml more simple.
| rules: | ||
| - apiGroups: | ||
| - "" | ||
| resources: |
There was a problem hiding this comment.
Are we really need to watch so many resource?
There was a problem hiding this comment.
Here are the resources we may deal with at the moment.
pods/endpoints/services/replicationcontrollers/replicationcontrollers/scale is needed by now.
configmaps/persistentvolumeclaims can be used to do some enhancements in future.
There was a problem hiding this comment.
Although I can't imagine what a controller watch so many resources can do, but if you are sure they are needed, so be it
charts/base/templates/rbac.yaml
Outdated
| metadata: | ||
| name: {{ .Values.rbac.serviceAccount }} | ||
| labels: | ||
| {{- include "base.labels" . | nindent 4 }} |
There was a problem hiding this comment.
Do we really need user to control the labels and annotation? This will complicate values.yaml
There was a problem hiding this comment.
labels are not necessary here, we can remove it.
| volumeMounts: | ||
| - mountPath: /ingress-apisix/conf | ||
| name: configuration | ||
| {{- with .Values.ingressController.nodeSelector }} |
There was a problem hiding this comment.
nodeSelector and affinity is similar, just one of them is enough
| imagePullPolicy: {{ .Values.ingressController.image.pullPolicy }} | ||
| ports: | ||
| - name: http | ||
| containerPort: {{ (.Values.ingressController.config.httpListen | split ":")._1 }} |
There was a problem hiding this comment.
.Values.ingressController.config.httpListen is :8080, but liveness work at 80(http)?
There was a problem hiding this comment.
Where shows the 80?
The containerPort is refined from the httpListen.
There was a problem hiding this comment.
You use a IANA_SVC_NAME, http, to define livenessProbe's port, according here, the port is 80.
charts/ingress-apisix/values.yaml
Outdated
| serviceAccount: "apisix-view-serviceaccount" | ||
|
|
||
| service: | ||
| type: ClusterIP |
There was a problem hiding this comment.
This option can not work well, if we selected NodePort, we maybe need to specified node port, but here is no such parameter.
There was a problem hiding this comment.
OK, for now, the service of ingress controller doesn't have too much sense (only /healthz and Prometheus metrics), will hardcoded the service type to ClusterIP
| type: ClusterIP | ||
| ports: | ||
| - port: {{ .Values.ingressController.service.port }} | ||
| targetPort: http |
There was a problem hiding this comment.
Here also use the container's port name
Closing #130.