App Suite Stack Chart
This stack chart is a "full stack deployment" of App Suite. It has Istio support built in.
Istio
App Suite root path
This chart allows to configure the path under which App Suite becomes available. By default the path is /appsuite.
This can get configured with the following snippet in your values.yaml:
global:
appsuite:
appRoot: "/appsuite"
appRoot must start with a slash, but must not end with a slash.
Adding additional routes
It is possible to add additional services to the VirtualHost provided by App Suite. The additional service needs to be located in the same namespace as App Suite itself.
To rewrite an exact path match to another path use the following example:
appsuite:
istio:
virtualServices:
appsuite:
extraRoutes:
- name: name-of-extra-route
matchExact:
- "/api/new-service"
rewrite: "/"
destinationHost: "new-service"
To rewrite a path prefix to another path use the following example:
appsuite:
istio:
virtualServices:
appsuite:
extraRoutes:
- name: name-of-extra-route
matchPrefix:
- "/api/new-service"
rewrite: "/"
destinationHost: "new-service"
To rewrite the path based on a regular expressione use the following example:
appsuite:
istio:
virtualServices:
appsuite:
extraRoutes:
- name: name-of-extra-route
matchRegex:
- "^/new-service/(.*)$"
rewriteMatchRegex: "^/new-service/(.*)/api/(.*)$"
rewriteRegex: "/\1/\2"
destinationHost: "new-service"
By default, the Helm release name will be added to the route's destinationHost. This might not be necessary for services that have not been deployed with a release name prefix. Therefore, the addition of the release name can be disabled with addReleaseName: false:
appsuite:
istio:
virtualServices:
appsuite:
extraRoutes:
- name: name-of-extra-route
matchRegex:
- "^/new-service/(.*)$"
rewriteMatchRegex: "^/new-service/(.*)/api/(.*)$"
rewriteRegex: "/\1/\2"
destinationHost: "new-service"
addReleaseName: false
Security headers
To enable security headers to configure referrer policy and transport security:
appsuite:
istio:
securityHeaders:
enabled: true
Destination rules
The chart renders one Istio DestinationRule per core-mw service. Each keeps a user or client on one core-mw pod through a consistent hash: a cookie for appsuite (http-api), the Authorization header for dav (sync) and businessmobility. Per block, these keys are passed through to Istio:
| Key | Rendered into |
|---|---|
istio.destinationRules.<block>.tls | trafficPolicy.tls |
istio.destinationRules.<block>.outlierDetection | trafficPolicy.outlierDetection |
istio.destinationRules.<block>.localityLbSetting | trafficPolicy.loadBalancer.localityLbSetting |
Warning: setting outlierDetection switches on Istio's locality failover for that host, since it is enabled mesh-wide by default. Each ingress gateway then hashes only over the core-mw pods in its own zone. If gateways in several zones serve one user, the user is spread across pods and per-pod state such as mail compose (uploads, inline images) breaks. Disable locality load balancing for the rule in that case:
appsuite:
istio:
destinationRules:
appsuite:
outlierDetection:
consecutive5xxErrors: 5
interval: 5s
baseEjectionTime: 30s
maxEjectionPercent: 100
localityLbSetting:
enabled: false
Optional dependencies
Redis
By default, core-ui-middleware is configured to start one pod. When you plan to scale it to a higher value, you will need Redis. This allows all core-ui-middleware pods to communicate and synchronize with each other.
The following configuration snippet can be used:
appsuite:
core-ui-middleware:
redis:
enabled: true
host: appsuite-redis-master
redis:
enabled: true
architecture: standalone
auth:
enabled: false
master:
persistence:
enabled: false
For this usecase its recommened not to scale Redis itself, as this might slow down core-ui-middleware due to additional overhead.