-
rf-version-info ConfigMap: Need information on data that is needed to populate this ConfigMap for deployments. TODO comments are inchart/charts/<microservice>/templates/deployment.yaml
- Configmap info was provided by RF, was implemented in parent chart via the
configmap.yaml
and theresources/.env
files, this may not be the best way to go about this, especially since the.env
file contains what should be a secret, but it works for now.
- Configmap info was provided by RF, was implemented in parent chart via the
-
Commented out
nodeSelector
field inchart/charts/backend/values.yaml
-
Created two PersistentVolumeClaims at
chart/charts/iso-master/templates/pvc.yaml
-
Redis is using the
latest
image tag - we should use an explicit Big Bang redis package -
HPA's will require the existence of a metrics server on the target cluster.
-
Will need real-world resource consumption by each service for realistic requests/limits in the templates
- backend
- /api/v1/.*
- port 80
- frontrow
- /.*
- /cli/.*
- port 80
- iso-master
- /iso-master/.*
- port 80
- keycloak
- /auth/.*
- port 80
- rf-scan
- /rf-scan/.*
- port 80
- rfapi
- rfapi seems to have some special sauce, here's the value definition of ingress
ingress:
websocket:
enabled: true
className: rf-ingress-class
annotations:
nginx.ingress.kubernetes.io/proxy-read-timeout: "3600"
nginx.ingress.kubernetes.io/proxy-send-timeout: "3600"
nginx.ingress.kubernetes.io/configuration-snippet: |
proxy_set_header Upgrade "websocket";
proxy_set_header Connection "Upgrade";
nginx.ingress.kubernetes.io/server-snippets: |
location /rfapi/events {
proxy_set_header X-Forwarded-Host $http_host;
proxy_set_header X-Forwarded-Proto $scheme;
proxy_set_header X-Forwarded-For $remote_addr;
proxy_set_header Host $host;
}
hosts:
- host:
paths:
- path: /rfapi/events
pathType: ImplementationSpecific
tls: []
http:
enabled: true
className: rf-ingress-class
annotations:
nginx.ingress.kubernetes.io/use-regex: "true"
hosts:
- host:
paths:
- path: /rfapi
pathType: ImplementationSpecific
tls: []
- rfpubsub
- /rfpubsub
- port 80
- code below is a snippet from values pertaining to ingress, prob have to implement this via istio
annotations:
nginx.ingress.kubernetes.io/proxy-read-timeout: "3600"
nginx.ingress.kubernetes.io/proxy-send-timeout: "3600"
nginx.ingress.kubernetes.io/configuration-snippet: |
proxy_set_header Upgrade "websocket";
proxy_set_header Connection "Upgrade";
nginx.ingress.kubernetes.io/server-snippets: |
location /rfpubsub {
proxy_set_header X-Forwarded-Host $http_host;
proxy_set_header X-Forwarded-Proto $scheme;
proxy_set_header X-Forwarded-For $remote_addr;
proxy_set_header Host $host;
}
- runner
- not sure if runner actually needs ingress or not, TODO
- /runner/.*
- port 80
- Uses a
latest
image tag - will need an explicit version - Mounts some custom hostPath directories - will require background information (IE -
/opt/rapidfort/data/aggregator/
named "raid-volume") - No service resource - how does the architecture utilize this service? data/processing?
- Uses a "stable" image tag - will need an explicit version
- Resource requests provided
- Mounts some custom hostPath directories - will require background information (IE -
/mnt/s3data
) KEYPAIR_NAME
environment variable is hardcoded- References a secret
rf-connection
that is not generated from a chart - HPA will require a metrics server on the target cluster
- Uses a "stable" image tag - will need an explicit version
- Will require resource declarations
- No persistent volume mounts
- Requires privileges
- Mounts
hostPath
andpersistentVolumes
- References a secret
rf-connection
that is not generated from a chart - Hardcoded environment variables in the deployment
- Exposes traffic externally
- initContainer has a hardcoded image
- Dependency on the ability to pull/run the rfstub apline base image?
- Uses
latest
image tag - will need an explicit version - Provisions an ingress for external access - is this required?
- Creates two ingresses for a single service (websocket/http) - pertinent if creating virtualservices
- HPA will require a metrics server on the target cluster
API_KEY
is hardcoded in the deployment.yaml - should consume value from values- Deployment/Values references a secret through
secretName
for which there is no secret resource created - No persistent data requirements
- Provisions an ingress for external access - is this required?
- Uses explicit
1.0.1
image tag
- uses
hostPath
for persistent data - uses
latest
image tag - HPA will require a metrics server on the target cluster
RF_RFVDB_REDIS_DB
environment variable is hardcoded in the deployment- Dependency on access to a redis service
- Deployment uses
hostPath
for host directory access / container runtime socket access. - Exposed via ingress to allow external traffic (CLI or other)
- Will require a virtual service
- Image uses a
latest
tag - will require an explicit tag to pass policies - Dependency on the backend service available
- Keycloak -
15.1.1
is not an available image version on registry1. Likely need to check compatibility with an updated image through the use of a bigbang keycloak package. - Would integration work with the latest keycloak release (v1.17.x)
- Worth pulling the existing bigbang package
- Uses the
latest
tag - Worth pulling the existing bigbang package
- Noted dependency to the rfvdb service
- Only defined as a dependency, no local package provided
- Document this dependency explicitly
- Entertain the option of using the big-bang mysql package for local dev / on-premise deployments