By name: analyze-sync-logs description: Generate a human-readable report from Airbyte sync log(s). Use when analyzing connector sync failures, comparing sync runs, or debugging connector issues. allowed-tools: Bash(python3:*)
Analyze the provided Airbyte sync log file(s) and present a concise diagnostic report.
Arguments: $ARGUMENTS (one or more log file paths)
Steps:
- For each log file, first determine the absolute path of the directory containing this SKILL.md file (call it SKILL_DIR), then run:
python3 $SKILL_DIR/scripts/analyze-sync-logs.py <file> - Parse the JSON output from each run
- Present a report for each log covering:
- Sync status, duration, start/end times
- Source and destination versions and images
- Key source config values that are present (e.g. url, cutoff_days, start_date, bucket_id/bucket_total, page_size)
- Key destination config values that are present (e.g. origin, edition, graph, graphql_api)
- Catalog: number of streams, list stream names with sync modes
- Records per stream (source reads) with total
- Destination stats: total read/processed/written/skipped/errored
- Top destination models by write count
- Errors and warnings (full messages)
- If multiple logs are provided, add a comparison section:
- Flag any differences in versions, config, or catalog
- Show record count changes per stream (delta and direction)
- Show model write count changes (top movers)
- Highlight anything that could explain failures or anomalies (e.g. large record count swings, new errors, status changes)
- End with a brief "Key observations" section calling out anything notable