Automatically Detect the Top Performance & Scalability Issues in Distributed
summary
Claims automated detection of common distributed anti-patterns across 100+ teams in 2019 and names the offenders: N+1 call & query, payload flood, tight coupling, bad timeouts/retries/backoff. Shows concrete failures — one search triggers 26,000 DB calls; another pipeline ships whole 10MB documents and then reduces network cost by >50% after refactor. Ships an event-driven Keptn extension that pulls Prometheus or Dynatrace metrics, grades SLIs/SLOs into a 0–100 score in CI/CD, and runs/remediates workflows with reevaluation until success.
AI-written summary · about
chapters (16)
- 0:00Intro
- 1:16Distributed Trace Example from StepStone (AWS Summit Berlin 2019)
- 5:36In distributed architectures we need to answer: Who is depending on me? What is the risk of change?
- 8:15N+1 Call Pattern across distributed "Product Service"
- 9:15LANDBAY
- 11:23Cascading N+1 Query Pattern: This is a single End-2-End Distributed Trace
- 12:18Payload Flood: "Doc Creation" sequential across distributed services
- 13:12Refactor: Only send relevant data to specialized services
- 20:03Keptn from 10000ft: Declarative Workflows + Event-Triggered Actions
- 21:58Root Cause: Lengthy manual approval in existing delivery pipelines
- 22:41Inspired by Dynatrace's internal „Performance Signature as Code
- 23:50SLI/SLO-based evaluation implementation in Keptn
- 25:33Solution: Automate Approval through SLI/SLO-based Quality Gates
- 26:10Demo: Automated SLI/SLO Validation based on Dynatrace Dashboards
- 30:22Keptn-Closed-Loop Remediation with Keptn 0.7
- 33:08Automate Distributed Problem Detection & Remediation
related talks
52mPerformance & Stability Anti patterns
37mUsing SLOs for Continuous Performance Optimizations of Your K8s Workloads
20mUsing Monitoring to Speed Up Your DevOps Journey
37mBuilding a Metrics Optimised Pipeline (Full Automation)
31mAlerting for Distributed Systems—A Tale of Symptoms and Causes, Signals and Noise
25mImplementing Metrics-Driven DevOps: Why and How
27mDesigning Your Pipelines for Scalability and Resiliency
46mTroubleshooting Distributed Workloads ...