Learning Track
Log Analysis at Scale
Quickly slice bid/impression logs to find latency spikes, error signatures, and delivery blockers.
Why it matters
Most complex issues only show up in logs. TAMs who can query Parquet/JSON at scale resolve deals and SPO faster.
What to read
- Parquet/JSON formats; Athena/BigQuery basics
- Common RTB error codes; timeout patterns
- How to bucket latency (p50/p90/p99)
How to practice
- Analyze 500 logs; group by error code and latency bucket.
- Identify top 10 error signatures; build regex filters.
- Visualize latency histogram; write a 1-page playbook.
Outputs to keep
- queries.sql, latency-histogram.png, log-playbook.md.
References
- AWS Athena, GCP BigQuery docs
- IAB RTB error code lists
- Grafana/QuickSight basics for quick charts