What causes "failed to write chrome trace: {}" in xai-org/grok-build?

Check free disk space (`df -h`) on the output target and clear space

failed to write chrome trace: {}

Error message

failed to write chrome trace: {}

What it means

After creating the output file, generate_chrome_trace serializes the trace with serde_json::to_writer_pretty; any serialization/stream error is wrapped in this message. Because the trace value is a simple serde_json::json! literal, failures here are almost always I/O errors on the write side (disk full, closed/failed file handle) rather than JSON encode bugs.

Source

Thrown at crates/codegen/xai-grok-telemetry/src/instrumentation.rs:505

        });

        events.push(trace_event);
        seen += 1;
    }

    if seen == 0 {
        return Err(anyhow!("no timing events found in {:?}", input));
    }

    let trace = serde_json::json!({
        "displayTimeUnit": "ms",
        "traceEvents": events,
    });

    let mut output_file = std::fs::File::create(&output)
        .map_err(|err| anyhow!("failed to create chrome trace {:?}: {}", output, err))?;
    serde_json::to_writer_pretty(&mut output_file, &trace)
        .map_err(|err| anyhow!("failed to write chrome trace: {}", err))?;

    Ok(output)
}

pub fn finalize() -> Result<()> {
    let mode = mode();
    if mode == InstrumentationMode::Disabled {
        return Ok(());
    }

    drop_guard(LOG_GUARD.get());
    drop_guard(CHROME_GUARD.get());

    Ok(())
}

fn drop_guard<T>(guard: Option<&Mutex<Option<T>>>) {
    if let Some(lock) = guard

View on GitHub (pinned to bc7f02eddd)

Solutions

Check free disk space (`df -h`) on the output target and clear space
Retry after confirming the filesystem is healthy and writable
Verify no disk quota is exceeded in CI/container environments
If it persists, inspect `err.source()`/raw io error for the specific OS code

Example fix

// before
serde_json::to_writer_pretty(&mut output_file, &trace)
    .map_err(|err| anyhow!("failed to write chrome trace: {}", err))?;
// after
serde_json::to_writer_pretty(&mut output_file, &trace)
    .map_err(|err| anyhow!("failed to write chrome trace: {}", err))?;
output_file.sync_all()
    .map_err(|err| anyhow!("failed to flush chrome trace: {}", err))?;

Defensive patterns

Strategy: try-catch

Validate before calling

// check space before writing
let meta = output.parent().and_then(|d| fs2::available_space(d).ok());
if meta.map_or(false, |bytes| bytes < 1_000_000) { eprintln!("low disk space"); }

Try / catch

match generate_chrome_trace(opts) {
    Err(e) if e.to_string().contains("failed to write chrome trace") => {
        eprintln!("write failed mid-stream: {e}; check disk space and retry");
        let _ = std::fs::remove_file(&partial_output); // clean partial file
    }
    other => other,
}

Prevention

Monitor free disk space before large writes
Clean up partial output files on failure
Avoid writing traces to network/removable mounts
Retain the underlying io::Error source for diagnosis

When it happens

Trigger: Calling generate_chrome_trace where the write to the already-created output file fails mid-stream — typically ENOSPC (disk full), a broken pipe, or an I/O error from the underlying file handle.

Common situations: Generating a large trace on a nearly-full disk; writing to a network mount or removable drive that dropped; container/CI disk quota exhausted during the write.

Related errors

AI-assisted analysis of xai-org/grok-build@bc7f02eddd (2026-08-31). Data as JSON: /api/errors/340cc83b87286202. Report an issue: GitHub.