fix(helm): align values with ts-server-boilerplate server/telemetry split

[CL-SHLOMIKONCHA]

Question	Answer
Bug fix	✔
New feature	✖
Breaking change	✔
Deprecations	✖
Documentation	✖
Tests added	✖
Chore	✖

Related issues: MAPCO-11633, MapColonies/ts-server-boilerplate#476 (MAPCO-11622), raster-export umbrella TODOs (helm-charts-new → charts/raster-export/values.yaml, exporter-overseer block)

Further information:

Default values

Value	Before	After
image.pullPolicy	Always	IfNotPresent
configManagement.offlineMode	false	true
telemetry.logger.opentelemetryOptions.enabled	false (was env.openTelemetryOptions.enabled)	true
telemetry.tracing.enabled	false (was tracing.enabled)	true

These resolve the # TODO: remove after set this as the default value within service overrides on exporter-overseer.image.pullPolicy and exporter-overseer.configManagement.offlineMode in the raster-export umbrella.

Server / telemetry split (mirrors ts-server-boilerplate#476)

• env.port, env.targetPort, env.protocol, env.requestPayloadLimit, env.responseCompressionEnabled move to a server scope.
• env.logLevel, env.logPrettyPrintEnabled, env.openTelemetryOptions and the top-level tracing block are replaced by a telemetry scope: logger.level, logger.prettyPrint, logger.opentelemetryOptions {enabled, url, resourceAttributes}, tracing {enabled, url}.
• The remaining overseer keys stay under env unchanged: env.httpRetry, env.disableHttpClientLogs, env.shapefileReader.
• Global overrides move under global.telemetry with the same hierarchy. Only the tracing and OTLP logger urls are overridable from global and take precedence over chart values; enabled flags stay chart-level.
• The OTLP logger url is exported as OTLP_LOGGER_ENDPOINT_URL and resource attributes as OTEL_RESOURCE_ATTRIBUTES (key=value,key=value), the names the commonBoilerplateV3 schema and the OpenTelemetry SDK read. The previous OPENTELEMETRY_LOGGING_URL / OPENTELEMETRY_RESOURCE_ATTRIBUTES keys were not consumed. config/custom-environment-variables.json maps telemetry.logger.opentelemetryOptions from the same names.
• OPENTELEMETRY_LOGGING_ENABLED and TELEMETRY_TRACING_ENABLED are always exported with their value; urls only when enabled.
• common.tracing.merged and common.openTelemetryOptions.merged helpers removed; overseer.tracingUrl, overseer.opentelemetryLoggingUrl and overseer.otelResourceAttributes added.
• TELEMETRY_METRICS_BUCKETS (from mclabels.prometheus.buckets) is untouched; the chart has no separate metrics scope.

Breaking for consumers: umbrella charts and site-values setting env.port, env.targetPort, env.protocol, env.requestPayloadLimit, env.responseCompressionEnabled → server.*; env.logLevel → telemetry.logger.level; env.logPrettyPrintEnabled → telemetry.logger.prettyPrint; env.openTelemetryOptions.* → telemetry.logger.opentelemetryOptions.*; tracing.* → telemetry.tracing.*; global.tracing.url → global.telemetry.tracing.url; global.openTelemetryOptions.url → global.telemetry.logger.opentelemetryOptions.url. Known spots: raster-export exporter-overseer.image.pullPolicy and exporter-overseer.configManagement.offlineMode can be dropped (now defaults); the commented # env: logLevel: info in site-values raster/export/{dev,qa}.yaml would become telemetry.logger.level if uncommented.

Note: helm lint with bare chart defaults fails on master and on this branch alike, because deployment.yaml reads storage.fs.ingestionSourcePvc.enabled which only site-values provides (global.storage.fs.ingestionSourcePvc.enabled: false). Rendering with that value set lints clean. Not changed here.

🤖 Generated with Claude Code