Cortex Configuration

caution

This Configuration is for Cortex dependency image conduktor/conduktor-console-cortex

This image is exclusively configured through Environment Variables.

The only required property is CDK_CONSOLE-URL, everything else is related to storage for the metrics.

By default, data will be stored in /var/conduktor/monitoring inside the running image. You can mount a volume on this folder to keep metrics data between updates. Otherwise, you can use the storage parameters described below to store the data using either s3, gcs, azure or swift

Environment Variable	Description	Mandatory	Type	Default	Since
CDK_CONSOLEURL	Console URL and port (example: "http://conduktor-console:8080").	true	string	∅	1.18.0
CDK_SCRAPER_SKIPSSLCHECK	Disable TLS check when scraping metrics from Console.	false	bool	false	1.18.2
CDK_SCRAPER_CAFILE	Path to CA certificate file inside the container to perform TLS check when scraping metrics from Console.	false	string	∅	1.18.2
S3
CDK_MONITORING_STORAGE_S3_ENDPOINT	S3 storage endpoint.	false	string	∅	1.18.0
CDK_MONITORING_STORAGE_S3_REGION	S3 storage region.	false	string	∅	1.18.0
CDK_MONITORING_STORAGE_S3_BUCKET	S3 storage bucket name.	true	string	∅	1.18.0
CDK_MONITORING_STORAGE_S3_INSECURE	S3 storage SSL/TLS check flag.	false	bool	false	1.18.0
CDK_MONITORING_STORAGE_S3_ACCESSKEYID	Access key ID of an AWS IAM identity for Monitoring to upload logs to S3. If set to false, and if you are running on a Kubernetes deployment, Monitoring can attempt to leverage AWS IRSA for the Pod’s service account when connecting to S3.	false	string	∅	1.18.0
CDK_MONITORING_STORAGE_S3_SECRETACCESSKEY	Secret access key of an AWS IAM identity for Monitoring to upload logs to S3. If set to false, and if you are running on a Kubernetes deployment, Monitoring can attempt to leverage AWS IRSA for the Pod’s service account when connecting to S3.	false	string	∅	1.18.0
GCS
CDK_MONITORING_STORAGE_GCS_BUCKETNAME	GCS storage bucket name.	true	string	∅	1.18.0
CDK_MONITORING_STORAGE_GCS_SERVICEACCOUNT	GCS storage service account JSON content.	true	string	∅	1.18.0
Azure
CDK_MONITORING_STORAGE_AZURE_ACCOUNTNAME	Azure storage account name.	true	string	∅	1.18.0
CDK_MONITORING_STORAGE_AZURE_ACCOUNTKEY	Azure storage account key.	true	string	∅	1.18.0
CDK_MONITORING_STORAGE_AZURE_CONTAINERNAME	Azure storage container name.	true	string	∅	1.18.0
CDK_MONITORING_STORAGE_AZURE_ENDPOINTSUFFIX	Azure storage endpoint suffix.	false	string	blob.core.windows.net	1.18.0
Swift
CDK_MONITORING_STORAGE_SWIFT_AUTHURL	Swift storage authentication URL.	true	string	∅	1.18.0
CDK_MONITORING_STORAGE_SWIFT_PASSWORD	Swift storage user password.	true	string	∅	1.18.0
CDK_MONITORING_STORAGE_SWIFT_CONTAINERNAME	Swift storage container name.	true	string	∅	1.18.0
CDK_MONITORING_STORAGE_SWIFT_USERID	Swift storage user ID.	false	string	∅	1.18.0
CDK_MONITORING_STORAGE_SWIFT_USERNAME	Swift storage user name.	false	string	∅	1.18.0
CDK_MONITORING_STORAGE_SWIFT_USERDOMAINNAME	Swift storage user domain name.	false	string	∅	1.18.0
CDK_MONITORING_STORAGE_SWIFT_USERDOMAINID	Swift storage user domain ID.	false	string	∅	1.18.0
CDK_MONITORING_STORAGE_SWIFT_DOMAINID	Swift storage user domain ID.	false	string	∅	1.18.0
CDK_MONITORING_STORAGE_SWIFT_DOMAINNAME	Swift storage user domain name.	false	string	∅	1.18.0
CDK_MONITORING_STORAGE_SWIFT_PROJECTID	Swift storage project ID.	false	string	∅	1.18.0
CDK_MONITORING_STORAGE_SWIFT_REGIONNAME	Swift storage region name.	false	string	∅	1.18.0
Logs
CORTEX_ROOT_LOG_LEVEL	Cortex log level.	false	string	info	1.18.0
CORTEX_ALERT_ROOT_LOG_LEVEL	Alert manager log level.	false	string	info	1.18.0
PROMETHEUS_ROOT_LOG_LEVEL	Prometheus log level.	false	string	info	1.18.0

Example configuration

In a docker compose it may look like the following:

version: '3.8'
services:
  conduktor-console:
    image: conduktor/conduktor-console
    ports:
      - "8080:8080"
    environment: 
      CDK_MONITORING_CORTEX-URL: http://conduktor-monitoring:9009/
      CDK_MONITORING_ALERT-MANAGER-URL: http://conduktor-monitoring:9010/
      CDK_MONITORING_CALLBACK-URL: http://conduktor-console:8080/monitoring/api/
      CDK_MONITORING_NOTIFICATIONS-CALLBACK-URL: http://localhost:8080
  conduktor-monitoring:
    image: conduktor/conduktor-console-cortex
    ports:
      - "9009:9009" # cortex api
      - "9010:9010" # alertmanager api
      - "9090:9090" # prometheus api
    environment:
      CDK_CONSOLE-URL: "http://conduktor-console:8080"

Overriding Configuration

Cortex configuration can be overridden completely by mounting a YAML file into path /opt/override-configs/cortex.yaml. You can also change the path location using the CORTEX_OVERRIDE_CONFIG_FILE environment variable.
This is not currently available for Alert Manager and Prometheus.

For example, create a file cortex.yaml add in only your overrides:

limits:
  ingestion_rate: 50000
  max_series_per_metric: 100000

Mount to /opt/override-configs/cortex.yaml.
Spin up the container. Exec into the container and confirm the contents, replace 2 with the number of lines of override you wish to see, or remove grep to get the whole file:
cat /var/conduktor/configs/monitoring-cortex.yaml | grep limits -A2.

You should see a similar entry to the below in the opening logs:

INFO monitoring_entrypoint - Patch "/var/conduktor/configs/monitoring-cortex.yaml" configuration with "/opt/override-configs/cortex.yaml" fragment

Troubleshooting

No metrics in the monitoring page

Go to http://localhost:9090/targets to see Prometheus scraping target status.

If it fails, check that you can query metrics endpoint from conduktor-console-cortex container.

You might also have to configure CDK_SCRAPER_SKIPSSLCHECK or CDK_SCRAPER_CAFILE if conduktor-console is configured with TLS termination.

No Slack notification alerts

1. Follow the steps to configure Slack integration in the Integrations tab. You'll be asked to create a Slack App and to set OAuth2 authentication token on Console.
2. Don't forget to manually add Slack App bot to the channel integrations you want to use for alerts notifications.
3. Enable notifications in the Alerts tab, and select the same channel as previously.
4. Create some alerts.

If you still have issues with monitoring and alerting setup please contact our support team.

Endpoint Authentication

Monitoring is not designed to be interacted with through the API endpoints by end users, only Console. As such no ingress is available externally and you should not set one up as there is no authentication mechanism.