Overview
AgentMon gives security and platform teams runtime oversight of the AI agents already running inside their infrastructure. Claude Code, Cursor, Codex, OpenClaw, custom LangChain workers, and more. As autonomous agents multiply cross hosts and act with real credentials at machine speed, conventional monitoring easures performance and cost but cannot answer the question that matters in an udit: who did that, and can you prove it? AgentMon treats every agent as an actor with real authority. It continuously learns what safe agent behavior looks like, detects dangerous actions in real time, and elps keep your entire agentic network accountable without requiring specialized security expertise. Built on CLI, eBPF-based telemetry and OpenTelemetry-native instrumentation, AgentMon works with any OTLP-instrumented agent, with no proprietary SDKs and no vendor lock-in.
This BYOL Single AMI deploys AgentMon as a self-contained appliance in your own AWS account, keeping all agent telemetry and security-relevant evidence inside your environment and under your retention controls not in a third-party vendor cloud.
Highlights
- **Know what's running:** a fleet dashboard showing every agent's status, model, runtime, and activity; live trace and log streaming; an auto-discovered agent opology map showing which agent spawned what and which tools and services were invoked.
- **Catch problems early:** cost-spike alerts, reasoning-loop detection, runaway-agent protection, and failure grouping by root cause routed to Slack, PagerDuty, email, or webhooks to fit your existing incident workflow.
- **Stay secure:** automatic secret redaction strips API keys, tokens, and sensitive headers from stored telemetry; anomalous-behavior flagging surfaces suspicious prompts and unexpected permission escalations; a full, append-only, searchable audit trail records every agent run and tool call to support investigation and regulatory needs; detections export to your SIEM.
Details
Introducing multi-product solutions
You can now purchase comprehensive solutions tailored to use cases and industries.
Features and programs
Financing for AWS Marketplace purchases
Pricing
Vendor refund policy
As a BYOL product, no software charges are billed through AWS marketplace; infrastructure charges follow AWS's standard refund policy. License refunds per your Codenotary agreement. EULA: Standard Contract for AWS Marketplace (pre-vetted) unless legal supplies a Codenotary addendum.
How can we make this page better?
Legal
Vendor terms and conditions
Content disclaimer
Delivery details
64-bit (x86) Amazon Machine Image (AMI)
Amazon Machine Image (AMI)
An AMI is a virtual image that provides the information required to launch an instance. Amazon EC2 (Elastic Compute Cloud) instances are virtual servers on which you can run your applications and workloads, offering varying combinations of CPU, memory, storage, and networking resources. You can launch as many instances from as many different AMIs as you need.
Version release notes
This release delivers the full AgentMon platform as a self-contained appliance running on a single-node Kubernetes (k3s) stack:
AGENT FLEET VISIBILITY Fleet dashboard showing the status, model, runtime, and live activity of all agents; real-time trace and log streaming; and an auto-discovered topology map of agents, spawned sub-agents, and invoked tools. Works out of the box with Claude Code, Cursor, Codex, Copilot, Gemini CLI, Goose, and any OTLP-instrumented agent, no proprietary SDK required.
RUNTIME SECURITY Per-action identity attribution so no agent acts anonymously; prompt-injection and poisoned-instruction detection at ingress using built-in, language-aware detection packs; automatic secret redaction (API keys, tokens, and sensitive headers) before telemetry is written to disk; an append-only, searchable audit trail of every agent run and tool invocation; and detection export to your SIEM.
POLICY ENFORCEMENT Promote any detection to an enforcement rule (warn, require approval, or block) with a single click.
OPERATIONS & COST Cost-spike alerts, reasoning-loop detection, runaway-agent protection, and root-cause failure grouping delivered to Slack, PagerDuty, email, or webhooks; plus per-agent, per-model, and per-team spend tracking with budget forecasting and chargeback reporting.
PLATFORM Ubuntu 24.04 LTS and single-node k3s (v1.36), with all container images pre-baked and no internet pulls required at first boot; TLS provided by Caddy; CloudWatch agent integration; and a built-in diagnostics bundle (amon diag) for support. All telemetry remains within your AWS account and is governed by your retention policies. BYOL: Contact sales@codenotary.com for licensing.
Additional details
Usage instructions
-
LAUNCH - Start the AMI on an m5.xlarge instance (recommended) with the recommended security group (HTTPS 443). The appliance is fully self-contained: no internet pulls are needed at first boot.
-
FIRST BOOT - Allow about 5 minutes. The appliance prepares its data volume, starts single-node Kubernetes (k3s), and deploys AgentMon from container images baked into the AMI.
-
GET CREDENTIALS - Connect via AWS Systems Manager Session Manager (aws ssm start-session --target <instance-id>) or SSH as user "agentmon". Run: sudo tail -40 /var/log/agentmon-firstboot.log for the dashboard URL. The one-time bootstrap admin password is in /var/log/agentmon-firstboot-secret.log, and can be recovered any time with: sudo kubectl -n agentmon get secret agentmon-bootstrap-admin -o jsonpath='{.data.bootstrap-admin-password}' | base64 -d
-
SIGN IN - Open https://<public-ip-or-fqdn>/ and sign in as admin@local with the one-time password. Change it immediately under Profile -> My Account -> Password. The admin@local account remains your break-glass login.
-
INSTALL YOUR LICENCE - Under Settings -> Licensing, paste your .tok licence key and click Install (no restart needed). This is a BYOL product and no evaluation licence is pre-installed, contact sales@codenotary.com to obtain one. Licence-gated features such as single sign-on require the licence first.
-
CONNECT AGENTS - All agent traffic uses HTTPS on port 443: install the amon CLI and run "amon watch", or point any OTLP-instrumented agent at https://<your-appliance-fqdn>/ OTLP is routed internally, no extra ports to open. Create ingestion tokens in the dashboard.
SUPPORT - support@codenotary.com . Attach a diagnostics bundle (amon diag) to support requests.
Support
Vendor support
AgentMon is supported by Codenotary. Contact support@codenotary.com for installation, configuration, and operational issues business-hours response (CET/EST). A step-by-step deployment guide ships with the product, and the appliance includes a built-in diagnostics bundle (amon diag) you can attach to any support request. Licence purchase and renewal: sales@codenotary.com or
AWS infrastructure support
AWS Support is a one-on-one, fast-response support channel that is staffed 24x7x365 with experienced and technical support engineers. The service helps customers of all sizes and technical abilities to successfully utilize the products and features provided by Amazon Web Services.