K8s Telemetry MCP Server connects Amazon Q, Claude, and Bedrock agents directly to your cluster's observability stack — logs, metrics, traces, and AWS infrastructure — via a single Helm install.
Read-only access to every layer of your Kubernetes observability stack, structured for AI assistants.
Query and analyze pod logs via Loki. Pattern detection, error categorization, and actionable recommendations built in.
PromQL queries and cluster health via Prometheus. CPU, memory, restarts, network throughput, and SLO compliance.
Distributed trace search and retrieval via Grafana Tempo. Pinpoint slow requests and error traces across services.
Incident timelines, SLO checks, cost attribution, and alert enrichment — correlating data across all sources automatically.
Events, recent deployments, HPA scaling history, and node pressure — pulled directly from the Kubernetes API.
CloudTrail audit logs, ECR vulnerability scanning, RDS Performance Insights, and ElastiCache metrics.
"Why did the payment pod crash in the last hour?"
"Is the checkout service meeting its 99.9% SLO this week?"
"Which namespace is consuming the most resources?"
"Who deleted the payment-service deployment?"
"Which ECR images have critical vulnerabilities?"
"Why is my pod pending? Show me the Kubernetes events."
Available on AWS Marketplace. Pay with your existing AWS committed spend — charges appear on your monthly AWS invoice.
Business hours: Monday–Friday, 09:00–18:00 CET. All support requests handled via email and GitHub Issues.
| Tier | First Response | Target Resolution |
|---|---|---|
| Standard ($29/mo) | 5 business days | Best effort |
| Professional ($99/mo) | 2 business days | 5 business days |
| Enterprise ($299/mo) | 1 business day | 3 business days |
We build Kubernetes observability tooling for AI assistants.
Engineering teams spend too much time switching between dashboards, running manual queries, and piecing together context during incidents. We built K8s Telemetry MCP Server to eliminate that friction — letting your AI assistant do the querying while you focus on fixing the problem.
Our product implements the Model Context Protocol (MCP), an open standard that allows AI assistants like Amazon Q, Claude, and Bedrock agents to call external tools and retrieve structured data. Every tool is read-only and every response is sanitized before it reaches your AI.
Sales inquiries, support, or general questions — we respond promptly.