Hi everyone,
I’d like to share something we’ve been working on at Lynxmind over the past months — a prototype AI assistant for Checkmk that we’ve named Pardinus.
The initial version started as a Chrome Extension that overlays an AI chatbot on top of the Checkmk interface. While this allowed us to build and iterate quickly, it had obvious limitations: it required a specific browser, an installed extension, and it lived “outside” the Checkmk UI.
Because of that, we’re now moving towards a proper native integration inside Checkmk, which offers a cleaner experience and aligns much better with how users actually interact with the platform.
It’s also worth mentioning that there has already been interesting exploration in the community around AI + Checkmk, including work shared by @aeckstein. Our project follows a completely separate approach and codebase, but it’s great to see multiple perspectives emerging in the ecosystem.
Below is a summary of what the prototype already supports and what’s coming next.
Current Features
-
Context-aware AI assistance
You can ask questions about Checkmk, monitoring concepts, plugin development, rule configuration, alerts, and overall platform usage. -
Monitoring-data reasoning
You can ask Pardinus for monitoring data about a specific host, service, application or group — and it can analyse alerts, metrics and recent events to propose a possible RCA (root cause analysis) based on patterns, correlations or historical behaviour. -
Troubleshooting guidance
Helps with common issues such as agent installation and baking, check execution, perfdata, notifications, rulesets, SNMP, REST API usage, and more. -
User-permission–aware actions
Depending on your Checkmk role and permissions, the assistant will be able to guide or automate typical platform operations:
– creating folders
– adding or modifying hosts
– editing rulesets
– managing labels and host groups
– triggering REST API calls
Essentially, most actions that a user can already perform in the UI or via the REST API can be assisted or automated. -
Flexible backend LLM support
Pardinus is not tied to any specific model. You will be able to connect it to:
– your own private/self-hosted LLM
– an internal corporate model
– or a public LLM provider
This allows organisations to choose the model that best fits their security, governance and cost requirements. -
Code examples
Pardinus can generate examples for:
– Check plugins (legacy & new API)
– Special agents
– Local checks
– WATO/Ruleset definitions
– REST API queries
– Python or shell snippets for automation -
Session-based chat history
Useful when iterating on plugin questions, rule logic, or troubleshooting steps.
What we are working on now
We’re migrating the Chrome Extension into a Checkmk GUI plugin — meaning:
-
no browser extension needed
-
accessible directly from the sidebar or a dedicated view
-
cleaner user-permission handling
-
possibility of packaging it later for the Checkmk plugin directory
A set of screenshots and a short demo video will be shared soon.
Why the name “Pardinus”?
Lynx pardinus — the Iberian lynx — is one of the world’s rarest and most agile wild cats.
It felt like the perfect metaphor for an AI designed to be smart, precise, and deeply integrated with monitoring.
Looking for Community Feedback
Before going further, we’d love to hear from the community:
-
What use cases would you find most valuable?
-
What type of AI interactions would help your daily operations?
-
Are there specific pages or workflows where an assistant like this would be most useful?
-
Would you prefer a snapin, a dedicated page, or both?
All feedback is welcome — this is being built for the Checkmk community, and we want it to be genuinely useful. You can address it here or by email checkmk@lynxmind.com
Thanks,
