Add Opentelemetry support to cmsmon-py base image

docker/cmsmon-py/Dockerfile

@@ -12,6 +12,13 @@ RUN apt-get update && apt-get install -y --no-install-recommends \
     krb5-user \
     && rm -rf /var/lib/apt/lists/*
 
-RUN pip install --no-cache-dir click pandas rucio-clients schema stomp.py==7.0.0
+RUN pip install --no-cache-dir \
+    click \
+    opentelemetry-exporter-otlp~=1.39.0 \
+    pandas \
+    rucio-clients \
+    schema \
+    stomp.py==7.0.0
 
+COPY helpers/ $WDIR/helpers/
 COPY src/ ./

docker/cmsmon-py/README.md

@@ -1,6 +1,6 @@
 # CMS Monitoring Python base image
 
-Dockerfile for generating the Base image for most non-Spark Python services in CMS Monitoring infrastructure.
+Dockerfile for generating the base image for most non-Spark Python services in CMS Monitoring infrastructure.
 
 ## Characteristics
 
@@ -15,6 +15,7 @@ Dockerfile for generating the Base image for most non-Spark Python services in C
 ### Pre-installed Packages
 
 - `click` - Command line interface creation toolkit
+- `opentelemetry-exporter-otlp~=1.39.0` - OTLP export for metrics, traces, and logs
 - `pandas` - Data manipulation and analysis library
 - `rucio-clients` - CMS data management system client
 - `schema` - Data validation library
@@ -23,3 +24,33 @@ Dockerfile for generating the Base image for most non-Spark Python services in C
 ## Usage
 
 This image functions as a base image for CMS Monitoring services that require Python but are not Spark-related. It is designed to be extended rather than run directly, and should be used as a base layer in Docker images that contain the specific scripts and code for individual services.
+
+## OpenTelemetry in child images
+
+Child images can send metrics, traces, and logs to the in-cluster OpenTelemetry Collector via `helpers/otel_setup.py`. Instrumentation is opt-in: set `OTEL_ENABLED=true` and `OTEL_SERVICE_NAME` in the workload environment (and `OTEL_EXPORTER_OTLP_ENDPOINT` if not using the default `opentelemetry-collector.opentelemetry.svc.cluster.local:4317`).
+
+Import `otel_setup` at startup (before `logging.getLogger()`). It configures stdout logging for `kubectl logs` and, when `OTEL_ENABLED=true`, also exports logs to the collector. Jobs should use `logging` (not `print`):
+
+```python
+import logging
+
+import helpers.otel_setup  # noqa: F401 — configures stdout + optional OTLP
+from helpers.otel_setup import global_meter, shutdown_opentelemetry, trace_span
+
+logger = logging.getLogger(__name__)
+
+@trace_span("my_job_step")
+def run_job():
+    logger.info("processing started")
+    global_meter.create_counter("jobs_processed").add(1)
+
+if __name__ == "__main__":
+    try:
+        run_job()
+    finally:
+        shutdown_opentelemetry()
+```
+
+See `helpers/otel_setup.py` for the full list of supported environment variables.
+
+Stomp loggers are set to `WARNING` on import (independent of `OTEL_ENABLED`). Re-call `configure_stomp_log_levels()` after `StompAMQ` init if needed.