Blog

Failure modes, root cause analysis, and the parts of a data pipeline that nobody has written down.

• 2026-11-04

  ETL Job Succeeded but No Data Loaded? 9 Causes and How to Find Yours

  Your ETL job reports success and the target table is empty or short. Nine causes, the check that confirms each, and how to make success mean something.

  etl job succeeded but no data loadedsilent data lossdata pipeline debugging

• 2026-11-04

  Google Cloud Logging Queries for Debugging Production Incidents

  Google Cloud Logging query patterns for incidents: structured logs, operator gotchas, comparing a bad period with a good one, and SQL over log buckets.

  Google Cloud Logging queryLogs ExplorerCloud Run

• 2026-11-04

  Hypothesis-Driven Debugging: A Field Guide for Production Incidents

  Hypothesis-driven debugging for production incidents: write rival explanations, predict what each implies, and run the cheapest test that tells them apart.

  hypothesis-driven debuggingdebuggingroot cause analysis

• 2026-11-04

  Incident Postmortem Template: Evidence, Rejected Hypotheses and Owned Actions

  A copyable incident postmortem template for outages, failed deploys, integration failures and data incidents, built around evidence and what you learned.

  incident postmortem templatepostmortemincident management

• 2026-11-04

  Observability vs Root Cause Analysis: What Each One Answers

  Observability vs root cause analysis: what detection and investigation each answer, where evidence lives, where they overlap, and how to run them together.

  observability vs root cause analysisobservabilityroot cause analysis

• 2026-11-04

  How to Reduce MTTR by Shortening Investigation, Not Gaming the Metric

  How to reduce MTTR: split resolution into phases, report medians not means, and shorten investigation, the phase where most controllable time is lost.

  reduce MTTRMTTRincident investigation

• 2026-11-04

  Root Cause Analysis for Software and Data Incidents: A Practical Guide

  A practical root cause analysis method for any production system: state the symptom, find where behaviour diverged, compare with a baseline, test rivals.

  root cause analysisincident investigationdebugging

• 2026-11-04

  Scheduled Job Troubleshooting: When a Job Fails Silently or Runs and Does Nothing

  How to troubleshoot scheduled jobs that fail silently or run and do nothing, with checks for cron, systemd, Kubernetes, Cloud Run, Airflow and SQL Agent.

  scheduled job troubleshootingcron job not runningKubernetes CronJob

• 2026-10-18

  Your pipeline succeeded and the rows are missing

  The worst ETL failures don't throw. They complete, report success and quietly drop rows nobody is counting. How silent row loss happens and how to find it.

  pipeline succeeded but rows missingsilent data lossrow count reconciliation