Blender CLI Logging and Render Farm Log Integration: Adding Timestamps for Failure Diagnosis

Blender's native --log option captures standard output to a file but does not provide internal timestamps for individual log entries. To add reliable time metadata or structured data without modifying the core logging subsystem, you must use a shell-level wrapper to intercept and augment the output stream.

Mechanisms for Adding Timestamps

Since Blender does not offer a built-in flag for ISO-8601 timestamps, the most effective approach is piping the output through a utility that prepends the current system time to each line. This ensures that every event is stamped upon arrival at the log file, allowing for correlation with external monitoring tools.

Shell-Based Timestamping

Depending on the available utilities on your render nodes, you can use the following methods. Note that 2>&1 is required to ensure that both standard output and error streams are timestamped:

• Using ts (moreutils): If installed, ts is the most direct tool for this purpose.
  blender -b input.blend -a | ts '[%Y-%m-%d %H:%M:%S]' > render.log 2>&1
• Using awk: A more portable alternative available on most Unix-like systems.
  blender -b input.blend -a | awk '{ print strftime("[%Y-%m-%d %H:%M:%S] "), $0 }' > render.log 2>&1

Structured Data and Machine-Readability

To achieve a machine-readable format (such as JSON), a lightweight wrapper script (e.g., in Python) is necessary. This script would read stdin line-by-line and output a structured object containing the timestamp and the log message. This is the only reliable way to implement structured fields without altering Blender's source code.

Analysis of Verbosity and Log Growth

Regarding the use of --verbose and log management:

• The --verbose flag: This should not be used as a proxy for timing. While it increases the volume of debug information, it does not provide sequential numbering or temporal markers that can be reliably correlated with a system clock.
• Log Rotation: Because Blender does not natively manage log file size or rotation, this must be enforced centrally. On Linux nodes, logrotate can be configured to manage growth. In containerized environments (e.g., Kubernetes), rotation should be handled by the container runtime or the log aggregator (e.g., Fluentd or Promtail) to prevent node disk exhaustion.

Verification and Buffering

When implementing these pipes, verify the following to ensure log integrity:

1. Buffering: Shell piping can introduce buffering, where logs appear in bursts. To force line-buffering, consider prefixing the command with stdbuf -oL.
2. Stream Capture: Confirm that the log file contains both the render progress (stdout) and any Python or C++ exceptions (stderr).

Diagnostic Detail Required: Are these nodes running as standalone bare-metal installs or within a container orchestrator? The specific implementation of log rotation and the available shell utilities vary significantly between these environments.