COGNIGY.AI in k8s

This is the official Kubernetes (k8s) manifest repository for COGNIGY.AI. You will need these files in order to deploy Cognigy.AI on top of Kubernetes. For dedicated instructions on how you can use this repository, please have a look at our production documentation.

CHANGELOG

This changelog contains changes to the structure of this repository. You can find the changelog for COGNIGY.AI here

3.4.0

Modification of files

Since we officially added Kubernetes support with our 3.4.0 release of COGNIGY.AI, we changed the structure of the repository a bit. We renamed several things:

• service-nlp-2.0.0-.yaml was renamed to service-nlp-2-0-0-.yaml
• The volume-claim "stateful-mongo-server.yaml" was renamed to "mongo-server.yaml"
• The volume "stateful-mongo-server.yaml" was renamed to "mongo-server.yaml"

In addition, the did the following things:

• We added a new secret to the deployment service-analytics-odata so that it can use the new conversations ODATA API.
• We added the environment variable NLP_REDIS_LOCK=false to all nlp deployments.
• We split the GE nlp deployment into several languages that now all reside in seperate folders witin the core/nlp-deployments folder.
• We also changed the config-map languages.yaml to add Portugese support and change some settings for Arabic. You therefore need to re-apply this config-map.

New limits

We also changed the limits of all of the deployments for our 3.4.0 release based on monitoring results, as well as changed the livenessProbe.

Monitoring

We added cadvisor support for monitoring. To use this, you have to re-apply the config-map monitoring/config-maps/prometheus-config.yaml and deploy the daemonset monitoring/daemonsets/cadvisor.yaml.

Management UI

You also need to update the Management UI to be compatible with 3.4.0. To do this, re-apply the deployment management-ui/deployments/management-ui.yaml