maestrod-0.6.0

Added

• Optional Prometheus Operator ServiceMonitor via
  observability.metrics.serviceMonitor.

document-engine-8.4.0

Added

• Added observability.log.structuredFlatten, which renders the LOG_STRUCTURED_FLATTEN environment variable. Only applies when observability.log.structured is true. Defaults to true, matching the Document Engine 1.16.0 default.

Changed

• Updated Document Engine to 1.16.0.
• config.tileMaxScale now defaults to unlimited (unset). Previously the chart pinned TILE_MAX_SCALE=16. Document Engine 1.16.0 enforces the limit on GET, POST and HTTP/2 tile rendering, so leaving it unset matches the upstream default. To preserve the previous behavior, set config.tileMaxScale: 16 explicitly.

maestrod-0.5.0

First public release. Value-compatible with the internal 0.3.4 chart; the
following defaults changed — set them explicitly to preserve old behaviour:

image:
  repository: <account>.dkr.ecr.eu-west-1.amazonaws.com/maestrod  # now pspdfkit/maestrod
  tag: nightly                  # now empty (→ appVersion)
  pullPolicy: Always                                                   # now IfNotPresent
imagePullSecrets: <...>
podLabels: { component_name: maestrod }  # now {}
restartJob:
  registryAuthSecretName: "<...>" # now ""

Added

• /health HTTP defaults for startupProbe / livenessProbe / readinessProbe.
• NUTRIENT_SHOW_SCALAR / NATIVESDK_VISION_LOGS via ConfigMap with
  checksum/config rollout trigger.
• serviceAccount, autoscaling, podDisruptionBudget, deploymentAnnotations,
  topologySpreadConstraints, schedulerName, lifecycle, extra*,
  sidecars, initContainers.
• licenseSecret.name: "" skips the NUTRIENT_LICENSE_KEY env var.
• Generated README.md + values.schema.json; ci/ values; helm test probe.

ai-assistant-2.0.0

Changed

• Breaking: config.serviceConfiguration now uses the v2 service configuration schema (version: "2" with providers and models). The v1 schema (version: "1" with aiServices) is no longer accepted. v2 enables label-based model selection, multi-provider routing, and request-time provider:model overrides.

  To upgrade, replace any v1 config.serviceConfiguration with a v2 configuration, or set config.serviceConfigurationConfigMap to supply the file from an external ConfigMap. See the v2 configuration guide.

• The default config.serviceConfiguration is an empty v2 configuration (version: "2" with empty providers and models), replacing the previous v1 aiServices example.

• Added config.serviceConfiguration.providers and config.serviceConfiguration.models values for the v2 schema.

ai-assistant-1.2.1

Fixed

• cloudNativePG.Database and cloudNativePG.document-engine.Database resources now set the required spec.owner field. Without it the CloudNativePG admission webhook rejects the resources with Database.postgresql.cnpg.io ... is invalid: [spec.owner: Required value, ...], breaking install of the chart when cloudNativePG.enabled is true. owner is sourced from database.postgres.username (defaults to postgres, which exists in the cluster via clusterSpec.enableSuperuserAccess: true).

document-engine-8.3.1

Changed

• Changed the default config.tileMaxScale value from 2 to 16.

document-engine-8.3.0

Added

• Added config.tileMaxScale, which renders the TILE_MAX_SCALE environment variable.

document-engine-8.2.3

Added

• New clustering panels in the Document Engine Grafana dashboard

document-engine-8.2.2

Changed

• MinIO references replaced with Garage

document-engine-8.2.1

Changed

• Document Engine 1.15.1
• New options workerPoolMaxRestarts and workerPoolMaxSeconds are available to control the worker supervisor restart throttling.