crawl-pipeline-rca

Root-cause methodology for distributed pipelines (command → event → handler chains over Lambda + EventBridge + SQS), with the article crawl system as the primary application. Use when a state-machine row is stuck non-terminal (e.g. crawlStatus='pending', summaryStatus='pending'), when a canary or SLO times out with the row never transitioning, when one entry point works but another doesn't for the same input (e.g.…