Keplar: Memory function

Introduction

Keplar is Scoutflo's memory function that revolutionizes how teams handle alert management and incident response. By leveraging historical data, pattern recognition, and continuous learning, Keplar automatically assesses incoming alerts and instantly matches them with the most relevant playbooks, dramatically reducing response times and improving resolution accuracy.

What is Keplar?

Keplar acts as your team's intelligent incident response advisor, functioning like an experienced engineer who remembers every past incident and knows exactly which procedures work best for specific alert patterns. Instead of manually sifting through playbooks or relying on tribal knowledge, Keplar instantly connects alerts with proven resolution strategies.

Why Use Keplar?

• Instant Intelligence: Get immediate playbook recommendations the moment an alert arrives

• Continuous Learning: Keplar gets smarter with every incident, improving recommendations over time

• Knowledge Preservation: Capture and retain institutional knowledge that survives team changes

• Faster Resolution: Eliminate the guesswork and decision paralysis during critical incidents

• Consistency: Ensure the same high-quality response regardless of who's on-call

How It Works

1

An alert fires

Your monitoring system detects an issue, a Prometheus alert, an AWS CloudWatch alarm, or a Sentry error.

2

You ask Voyager

You describe the alert or paste the error message. For example: "I'm seeing a high memory alert on my production pods" or "Help me understand this database connection error."

3

Kepler finds the match

Behind the scenes, Voyager queries Kepler to find the Playbook that matches your specific issue.

4

You get deterministic guidance

Voyager presents the Playbook content, explaining what the alert means, what's at risk, and walking you through prioritized investigation steps.

5

You investigate with confidence

You follow the proven steps, knowing they're based on expert knowledge, not AI guesswork.

Supported Alert Types

• Infrastructure Alerts

  • CPU utilization spikes and performance degradation

  • Memory exhaustion and garbage collection issues

  • Disk space alerts and I/O bottlenecks

  • Network connectivity and latency problems

• Application Performance Alerts

  • Response time degradation and timeout issues

  • Error rate spikes and exception patterns

  • Throughput drops and capacity problems

  • Service dependency failures

Fallback Strategies

• Topology-Based Context

  • Analyzes system topology and service dependencies

  • Suggests relevant playbooks based on infrastructure relationships

  • Considers blast radius and potential impact areas

• Generic Troubleshooting Guidance

  • Falls back to proven general troubleshooting procedures

  • Provides framework for systematic investigation

  • Guides teams through structured diagnostic processes

Getting Started

Connect Your Monitoring Tools

Investigate with Voyager

Scoutflo 101