Add --compile-bytecode flag to uv pip install commands for faster container startup

[coderabbitai]

https://docs.astral.sh/uv/reference/cli/#uv-pip-install--compile-bytecode

Problem Description

Container startup performance is significantly degraded due to missing --compile-bytecode flag in uv pip install commands across all Dockerfiles. This causes Python modules to compile bytecode at runtime during container startup instead of pre-compiling during the Docker build phase.

Impact

• CI Test Failures: Containers repeatedly fail Kubernetes readiness/liveness probes with connection refused errors
• Slow Startup: Container startup hangs during jupyter elyra --generate-config due to slow Python module imports
• Non-deterministic Performance: Startup times vary significantly depending on runtime compilation overhead
• User Experience: Poor container startup reliability in production environments

Root Cause Analysis

Python modules without pre-compiled bytecode (.pyc files) trigger on-the-fly compilation during import, causing:

1. Significant startup delays when importing many modules (common in Jupyter environments)
2. Kubernetes probe timeouts (typically 5-10 seconds) being exceeded
3. Container restart loops in orchestrated environments

Affected Scope

35 Dockerfiles across the entire notebook ecosystem are affected:

Image Types:

• codeserver: ubi9-python-3.11, ubi9-python-3.12 (CPU variants)
• jupyter: datascience, minimal, pytorch, tensorflow, trustyai (CPU, CUDA, ROCm variants for both Python 3.11 and 3.12)
• runtimes: datascience, minimal, pytorch, tensorflow, rocm-pytorch, rocm-tensorflow (CPU, CUDA, ROCm variants)
• rstudio: c9s-python-3.11, rhel9-python-3.11 (CPU, CUDA variants)

Current Problematic Pattern:

uv pip install --strict --no-deps --no-cache --no-config --no-progress --verify-hashes --index-strategy=unsafe-best-match --requirements=./requirements.txt --build-constraints=./requirements.txt

Solution

Add --compile-bytecode flag to all uv pip install commands:

uv pip install --compile-bytecode --strict --no-deps --no-cache --no-config --no-progress --verify-hashes --index-strategy=unsafe-best-match --requirements=./requirements.txt --build-constraints=./requirements.txt

Benefits:

1. Pre-compiled Bytecode: .pyc files generated during Docker build, not at runtime
2. Faster Startup: Eliminates compilation overhead during container initialization
3. Deterministic Performance: Consistent startup times across deployments
4. Probe Reliability: Containers become ready within Kubernetes timeout windows
5. Better UX: Reliable container startup in production environments

Implementation Approach

1. Systematic Update: Add --compile-bytecode flag to all 35 affected Dockerfiles
2. Consistent Placement: Position flag consistently (after --compile-bytecode before other flags)
3. Verification Testing: Ensure CI tests pass with improved startup performance
4. Performance Validation: Measure startup time improvements

Acceptance Criteria

Core Requirements:

• All 35 Dockerfiles updated with --compile-bytecode flag
• Flag positioned consistently in command structure
• No regression in package installation functionality
• CI tests pass without probe timeout failures

Verification Steps:

• Container startup time measurements show improvement
• Kubernetes readiness/liveness probes succeed consistently
• Python module imports complete faster at runtime
• No build failures or dependency resolution issues

Testing Coverage:

• All affected image types (codeserver, jupyter, runtimes, rstudio)
• Both Python 3.11 and 3.12 variants
• CPU, CUDA, and ROCm accelerator configurations
• Cross-platform compatibility (amd64, arm64)

Context Links

• PR: RHOAIENG-21690: chore(Dockerfiles): augument micropipenv with uv for faster package installs #968 - "RHOAIENG-21690: chore(Dockerfiles): augment micropipenv with uv for faster package installs"
• Discussion: RHOAIENG-21690: chore(Dockerfiles): augument micropipenv with uv for faster package installs #968 (comment)
• CI Failure Example: https://github.com/jiridanek/notebooks/actions/runs/16784807126/job/47532526249

Related Documentation

• uv pip install --compile-bytecode documentation