> ## Documentation Index > Fetch the complete documentation index at: https://www.dataframer.ai/docs/llms.txt > Use this file to discover all available pages before exploring further. # Tracking > Pin the patterns that matter and watch how often they come back. Once a pattern proves it matters, move it from exploration to monitoring. **Tracking** shows pinned Findings you watch over time, the patterns you already know matter. Once a pattern is pinned, DataFramer re-checks every new trace against it and shows how often it comes back. **Slack alerts:** turn on Slack alerts for any pinned collection to get notified when a tracked pattern shows up in new traces. ## Root causes and suggested fixes Knowing that a failure keeps coming back is only half the answer. Open a tracked finding and run the analysis, and DataFramer reports the distinct root causes behind it — each one cited to the traces it was drawn from, with a suggested fix where the evidence supports one, down to the prompt edit. Run it again after you ship a fix to see whether that cause is gone. The analysis reads a sample of the finding's traces rather than every one, so most traces come back attributed to no cause. It only runs on findings marked as failure modes: new findings are marked for you as a Discovery groups them, and you can set or clear that mark yourself on any tracked finding. ## Next steps Route the patterns you track to a human reviewer