Overall Support Position
AtScale’s container technology is modeled for Kubernetes orchestration and installation is designed for Helm Chart deployment. Containerized AtScale is proprietary software and all services should be treated as part of its proprietary stack. While containers may be deployed using various methods and using various container orchestration tools, AtScale supports deployment using Kubernetes with AtScale engineered Helm Charts. Manual or custom deployment should be carefully considered and tested by the customer as it becomes the customer’s responsibility to ensure proper deployment, management/communication of services and on-going maintenance like upgrades. Additionally, AtScale Quality Assurance Testing will only be conducted with packaged deployment scripting.
An AtScale deployment using packaged install scripts (Helm Charts) is fully supported. Under supported deployment conditions, AtScale will provide full technical support for all services bundled within the delivered solution. Customizations can be (not exclusive list):
- Interchanging services (e.g. Using your Postgres instead of AtScale’s bundled Postgres)
- Calling the embedded APIs instead of using AtScale APIs (e.g. Managing users with Keycloak’s API instead of AtScale user management APIs)
- Fundamentally altering the AtScale containers
- Any change that would alter the behavior of the service or the communication between services such as interchanging load balancers
- Turning off / not installing a particular service
These customizations would not be recommended as AtScale cannot bear the responsibility of interoperability once these alterations are made. Such deployments can impact the logging AtScale captures making it very difficult to debug the system as a whole. In a customized deployment of the AtScale platform, AtScale Support can/will only be responsible for AtScale runtime services (engine, app and API) and no effort will be made to debug a customer’s scripts or interchanged/swapped service unless explicitly agreed upon and documented otherwise. AtScale support will provide limited best effort to any customer configuration that deviates from the reference configuration, but AtScale will not be responsible for the stability of the platform under customized deployment scenarios.
Unsupported Modifications
While not an exhaustive list, these scenarios are not recommended by AtScale as they will have a predictably negative impact on the AtScale platform and could render the system inoperable:
- Interchange an embedded service for what they perceive as a like-for-like service (e.g. swapping mySQL database for the embedded Postgres database).
- Combining AtScale services into a combined container.
If a customer requires support for an otherwise unsupported implementation, there must be an explicit request made in writing and agreed upon resources to support. This support can include tech support, professional services, engineering, etc. Agreements should be amended to reflect the exception.
Services Matrix
The matrix below outlines all services found in the AtScale containerized solution and their configurability.
Service | Tech Used | Configurable | Notes |
Engine | Proprietary | No | |
API | Proprietary | No | |
WebUI | Proprietary | No | |
Entitlement | Proprietary | No | |
Message Broker | Redis | No | |
Services Database | Postgres | No | |
Ingress Controller | Kong | No | |
Secrets Broker | Vault | Yes | AtScale can integrate with an external Vault instance |
Identity Broker | Keycloak | Yes | AtScale can/should be configured with an Identity Provider. |
Telemetry Collector | OpenTelemetry | Yes | Optional outputting telemetry to any standard collector (e.g. DataDog, New Relic, etc). |
Logging Agent | Grafana Promtail | No | |
Observability Dashboarding | Grafana | No | Any desired changes in monitoring would be conducted in the client’s monitoring platform. |
Metrics Storage | Grafana Mimir | No | |
Tracing Storage | Grafana Tempo | No | |
Log Storage | Grafana Loki | No | |
Object Store | MinIO | Yes | Optional deployment, can use client’s object storage |
Required Ecosystem Components
External Service | Notes |
Identity Provider | Microsoft Entra ID, Active Directory, Okta |
Code Repo | GitHub, GitLab |
Data Platform | Snowflake, Databricks, GBQ, Postgres, InterSystems IRIS |
Please check the AtScale Supported Tools and Platforms page in the documentation for the most current components & versions for each AtScale release.