4.67 - Happy New Year 🎉❄️¶

Removal of text-davinci-003 by OpenAI

Due to the sunset of the text-davinci-003 model by OpenAI on January 4, 2024, this model will no longer be available in Cognigy.AI.

Weaviate Upgrade to v1.22.5

Starting with Cognigy.AI release 4.68.0, Weaviate v1.19.0 will no longer be supported. If you have a Cognigy on-premises installation and are utilizing the Knowledge AI features, upgrade Weaviate to v1.22.5. For a smooth upgrade, refer to the guide.

Deprecation of Single Replica Redis

Starting from Cognigy.AI release v4.65, the Single Replica Redis setup is deprecated and replaced with Redis in High Availability (HA) Mode. For on-premises installations, before upgrading to Cognigy.AI v4.65, ensure that you migrate according to the Redis: Single Replica to Redis HA Migration.

4.67.1¶

Cognigy.AI¶

Released January 16^th, 2024

Bugfixes¶

Fixed the issue where the Flow would sometimes start at the wrong Node after a Go To Node

Cognigy Voice Gateway¶

Released January 10^th, 2024

Bugfixes¶

Fixed the issue where the recognizer label was not correctly identified, potentially leading to the use of incorrect speech credentials
Reverted changes made in version 4.67.0 for the Label list on the Application page

Infrastructure¶

Cognigy Voice Gateway¶

Add variables to the StatefulSet sbc-rtp to control the start and end ports for RTP:
- .Values.rtpEngine.startPort (default=40000)
- .Values.rtpEngine.endPort (default=60000)
Add an environment variable to the Drachtio server container in the feature-server StatefulSet and the sbc-sip StatefulSet to enable support for best-effort TLS:
- .Values.drachtio.supportBestEffortTls (default=false)

4.67.0¶

Cognigy.AI¶

Released January 9^th, 2024

Features¶

Added the Generic Settings section to the Voice Gateway Endpoint

Improvements¶

Added the Referred By option to the Refer transfer type in both the Call Failover and Call Events (specifically for Transfer Dial Error and Transfer Refer Error events) sections within the Voice Gateway Endpoint
Added the nova-2 Deepgram STT model
Added ElevenLabs as a Text-to-Speech (TTS) provider
Improved the Voice Gateway Endpoint by adding a setting to trim input.data for better transcripts
Added support to import the secondary locale of the package as the primary locale of the agent

Bugfixes¶

Fixed the memory leak issue in service-nlp-ner by replacing the axios package
Fixed the issue where LLM prompts in chat mode could fail if inputs contained no text
Fixed the issue where there were multiple entries in the Intent Trainer for the same Flow and input text
Fixed the issue where the UX of filter selection for the Case Node was misleading
Fixed the issue where the current year was not diplayed in the Billable Conversations filter on the Organisation Details page

Cognigy Insights¶

Released January 9^th, 2024

Improvements¶

Improved the UI of the Insights dashboards
Changed the dashboards' width to use all the available space on the screen

Bugfixes¶

Fixed the issue where the maximum message count was incorrect when the first digit is smaller than the minimum message count

Cognigy Live Agent¶

Features¶

Added the capability for human agents to create draft messages. If a reply or a private message is composed but not sent immediately, the message will be saved for 24 hours

Improvements¶

Renamed Live Agent Assist Bot to Copilot Bot

Cognigy Voice Gateway¶

Released January 9^th, 2024

Bugfixes¶

Fixed the issue where the default label in the UI was not applied
Fixed the issue where the selected language did not promptly map to the corresponding voices

Infrastructure¶

Cognigy Voice Gateway¶

Updated the helm.sh/hook annotations for the following Secret resources to "helm.sh/hook": "pre-install,pre-upgrade":
- voicegateway-call-recording-credentials
- voicegateway-drachtio-secret
- voicegateway-jwt
- voicegateway-mysql-password
Remove unused replicaCount from spec.Values for:
- rtpEngine
- rtpEngineSidecar
- sbcSipSidecar
With the release of 4.65.0, we introduced StatefulSets for sbc-sip and sbc-rtp. We recommend enabling StatefulMode as it improves the load balancing of traffic between the replicas.
- The required changes in values.yaml:
  - spec.Values.sbc.sip.useStatefulSet: true
  - spec.Values.sbc.rtp.useStatefulSet: true
```
sbc:
  sip:      
    useStatefulSet: true
  rtp:      
    useStatefulSet: true
```

With the release of 4.65.0, we introduced Redis High Availability. We recommend enabling redisHa:

The required changes in values.yaml:

cognigyEnv:
  FEATURE_REDIS_SENTINEL_MODE_ENABLED: "true"
  REDIS_SENTINEL_MASTER_NAME: "mymaster"
  REDIS_SENTINEL_HOST: "redis-ha-headless"
  REDIS_SENTINEL_PORT: "26379"

redisHa:
  enabled: true
metrics:
  serviceMonitor:
    enabled: true

With the release of 4.64.0, we disabled Grafana via a feature flag since we are using cognigy-monitoring-stack, which already includes Grafana. In the current release, we are removing all Grafana-related resources from the Helm charts:
- Service: grafana/service.yaml
- Deployment: grafana/deployment.yaml
- ConfigMap: grafana/config-map.yaml
- StorageClasses:
  - templates/gcp/grafana/storage-class.yaml
  - templates/azure/grafana/storage-class.yaml
  - templates/aws/grafana/storage-class.yaml

Version Compatibility Matrix¶

The version compatibility matrix was updated for the following Cognigy products: