dataops-self-healing-platform

Detect and recover from Airflow, Kafka, Kubernetes, and data quality failures.

14|1|Updated May 5, 2026
One-click install
npx skills add https://github.com/ivanshamaev/de-agent-skills --skill dataops-self-healing-platform
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: dataops-self-healing-platform
Source: https://github.com/ivanshamaev/de-agent-skills/tree/main/group_skills/infra_dataops_group_skills/dataops_self_healing_platform
Command: npx skills add https://github.com/ivanshamaev/de-agent-skills --skill dataops-self-healing-platform

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill reduces on-call workload by automatically recovering from common data platform failures before they escalate into prolonged data downtime or missing data.

Core Features & Use Cases

  • Self-healing Airflow orchestration: Auto-restart failed DAG runs with guardrails for idempotency and a circuit breaker to prevent retry storms.
  • Kafka consumer auto-recovery: Detect consumer lag/partition rebalance issues and apply automated actions such as scale-up, offset reset, or alerting.
  • Data reliability protections: Auto-quarantine partitions that fail DQ gates, detect missing partitions, and trigger automated backfill; include watchdog heartbeat monitoring and OOM-based Kubernetes scaling.

Quick Start

Ask an agent to set up the self_healing_watchdog workflow to auto-restart eligible failed Airflow DAG runs and quarantine partitions that fail DQ checks while monitoring Kafka lag and pipeline heartbeats.

Frequently Asked Questions about dataops-self-healing-platform

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automatically restart failed Airflow DAG runs without causing retry storms?

To implement self-healing for Airflow DAGs, set up an automated watchdog workflow that detects failed DAG runs and restarts them. It enforces idempotent recovery and uses circuit breaking to prevent retry storms from flapping pipelines.

What is a DQ quarantine pattern for data pipelines?

A DQ quarantine pattern isolates data partitions that fail data quality gates, preventing corrupted data from flowing downstream. It works alongside gap detection to trigger automated backfills for missing data partitions.

How do I resolve Kafka consumer lag and partition rebalance issues automatically?

You can resolve Kafka consumer lag automatically by deploying self-healing actions like consumer scale-up, offset reset, or targeted alerting. This detects lag and partition rebalance issues to recover consumers without manual intervention.

Can I use Kubernetes to auto-scale pods recovering from OOMKilled errors?

Yes, you can configure Kubernetes self-healing to detect OOMKilled containers and automatically apply scaling actions. This OOM-based scaling helps recover workloads and prevent future outages by adjusting resource allocations dynamically.

How do I detect missing partitions and trigger automated backfill in ETL pipelines?

To detect missing partitions and trigger automated backfill, implement a self-healing data reliability workflow. It monitors ETL pipeline heartbeats, identifies gaps in partition arrivals, and automatically initiates backfill processes to restore data.

When should I use circuit breaking for data pipelines?

You should use circuit breaking for data pipelines when retries become continuous or flapping, which can overwhelm downstream systems. It acts as a safeguard to halt automated recovery actions after exceeding guarded retry limits.