What problem does it solve?
Erigon's qa-* GitHub Actions workflows run for hours on self-hosted runners and produce huge annotated logs, result JSON files, and metric artifacts. This Skill teaches you how to triage a red QA - ... check efficiently: read the verdict from result-<chain>.json, grep the *** report lines, and map each failure mode (sync-time threshold, exec_steps_in_db, Deadline reached, RPC diff mismatches) to its actual root cause instead of dumping hundreds of megabytes of debug logs.
Core Features & Use Cases
- Verdict-first triage order: Start from
result-<chain>.json (outcome + reason), then the *** report lines, then the Erigon debug log and goroutine stack dump only when needed.
- Failure-mode decoding: Explains the three sync-test failure types (sync-time below 75%,
exec_steps_in_db over 300, Deadline reached), the exec-from-zero state-snapshot hash comparison, RPC integration diff mismatches, and RPC performance change-point detection.
- Workflow and artifact reference: Per-workflow tables covering runner labels, drivers, thresholds, artifact names, and MongoDB/Grafana history for regression comparison.
- Use Case: A PR shows a red
QA - Tip tracking check. Download the artifacts with gh run download, read outcome and reason from the result JSON, and determine in minutes whether it is a performance regression, a crash, or genuinely an infrastructure problem.
Quick Start
Ask the assistant to explain why the QA tip-tracking check failed on a specific GitHub Actions run of the erigon repository.