helm: align telemetry values and fix default values (MAPCO-11633)

[CL-SHLOMIKONCHA]

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

Related issues: MAPCO-11633, raster-serving umbrella chart TODOs (helm-charts-new → charts/raster-serving/values.yaml, site-values → raster/serving/*), MapColonies/ts-server-boilerplate#476 (MAPCO-11622)

Further information:

Defaults (resolve the raster-serving / site-values # TODO overrides for mapproxy)

Key	Before	After
imagePullPolicy	Always	IfNotPresent
db.schema	''	MapproxyConfig
mapproxy.db.nameOverride / mapproxy.db.schema	present	removed (never read; DB_SCHEMA comes from db / global.db)
tracing.enabled / global.tracing.*	false / merged from global	telemetry.tracing.enabled: true, url http://localhost:4317, samplingRatioDenominator: 1000
metrics.enabled / metrics.url / global.metrics.*	false, url required to enable	telemetry.metrics.enabled: true, no url

Telemetry values (aligned with the boilerplate telemetry scope)

• telemetry.tracing (enabled, url, samplingRatioDenominator) replaces the top-level tracing block and global.tracing. Only the url is overridable from global.telemetry.tracing.url and it takes precedence over the chart url; enabled stays chart-level. New helper mapproxy.tracingUrl; common.tracing.merged is removed.
• The mapproxy configmap exports TELEMETRY_TRACING_ENABLED (always) and TELEMETRY_TRACING_ENDPOINT (when enabled), both quoted; TELEMETRY_TRACING_SAMPLING_RATIO_DENOMINATOR is unchanged. The mapproxinator configmap exports TELEMETRY_TRACING_ENABLED (always) and TELEMETRY_TRACING_URL (when enabled). Both containers receive the same url value, as before.
• telemetry.metrics.enabled alone gates the uwsgi stats server (container env and mapProxyUwsgi.ini), the uwsgi-exporter sidecar and the mapproxinator TELEMETRY_METRICS_ENABLED flag. The former metrics.url was only echoed as TELEMETRY_METRICS_URL, which mapproxinator does not consume, so it is dropped together with global.metrics and common.metrics.merged. Prometheus scrape annotations keep coming from mclabels.prometheus (port 9117, already the default; that umbrella TODO is stale).
• mapproxy.env.telemetry.* (boto / http / sql / tileCache / traceDebug flags) is unchanged.

Fix: the init container gated REQUESTS_CA_BUNDLE on global.ca.enabled, which is not a defined value, so the bundle was never set; it now follows global.ca.secretName like the other containers. The wmts deployment guard on the same key could never fire and is removed.

Breaking for consumers: umbrella charts and site values setting top-level tracing.* / metrics.* or global.tracing.* / global.metrics.* must move to telemetry.tracing.*, telemetry.metrics.enabled and global.telemetry.tracing.url. In raster-serving/values.yaml: mapproxy.metrics (enabled + url) can be dropped (metrics default on, no url needed), mapproxy.imagePullPolicy and mapproxy.db.schema can be dropped (defaults now match), mapproxy.mclabels.prometheus.port: 9117 can be dropped (already the default), and mapproxy.mapproxy.imagePullPolicy / mapproxy.uwsgiExporter.imagePullPolicy were never read by the chart (all containers use the top-level imagePullPolicy) and can be dropped. In raster/serving/qa.yaml: mapproxy.imagePullPolicy and mapproxy.mapproxy.imagePullPolicy can be dropped. The nginx subchart's pull policy is owned by the nginx chart.

🤖 Generated with Claude Code