Conversational Operations
EN/PT
Language support
Pardinus Chat brings natural-language operations into Checkmk.
- Task, Report, and Wiki modes
- English and Portuguese interaction
AI Operations for Checkmk
Pardinus Suite Chat + RCA
AI capabilities built directly into Checkmk. Pardinus Chat supports natural-language operations. Pardinus RCA turns alerts into clear incident findings with next steps.
Critical Incident
DB01
CPU saturation at 98%
Context gathered before analysis so the result starts from system state, related signals, and operational history instead of guesswork.
17
Evidence items
1.84s
Pipeline time
0.87
Confidence
Chat
Natural-language operations for action, reporting, and learning in the tools teams already use.
RCA
Root-cause analysis with confidence, supporting context, and recommended actions for incident follow-up.
One AI operations layer, two products
Pardinus Chat and Pardinus RCA share the same integration and deployment model. Together they cover day-to-day interaction and deeper incident investigation.
Pardinus Chat
A conversational assistant that helps teams investigate issues, answer questions, and run routine actions in plain language.
- Task, Report, and Wiki modes
- English and Portuguese support
- Voice input, spoken replies, and per-user history
Pardinus RCA
A root-cause analysis workflow that turns raw alerts into clear incident reports with likely cause and next steps.
- One-click RCA from host and service actions
- Confidence scoring with alternatives
- Structured output for post-mortems and compliance
Explore the suite in motion.
See Pardinus in context
Pardinus stays embedded in the Checkmk workspace, giving operators a live operational view with immediate access to context, navigation, and follow-up actions.
API-first. Wire it into anything.
Every Pardinus capability is exposed through a versioned REST API. Trigger RCA from your incident pipeline, fetch reports into your ticketing tool, or attach evidence to collaboration channels without changing how teams already operate.
REST Endpoints
POST
/api/v1/rca/run Trigger an RCA job from an external workflow.
GET
/api/v1/rca/{job_id} Poll job status and retrieve results.
GET
/api/v1/rca/{job_id}/report.json Fetch structured output for downstream systems.
POST
/api/v1/chat/query Send conversational requests into Pardinus Chat.
Integrates with existing tooling
n8n
PagerDuty
ServiceNow
Jira
Slack
MS Teams
Webhooks
Your stack
Example request
curl -X POST https://checkmk.example/api/v1/rca/run \
-H "Authorization: Bearer $PARDINUS_TOKEN" \
-H "Content-Type: application/json" \
-d '[object Object]'202 Accepted. Job queued and ready for callback or polling.
Workflow example
Auto-attach RCA to every new incident
Setup time
~15 min
Integration mode
Webhook
Auth
Bearer
Checkmk alert
Automation trigger
POST /api/v1/rca/run
Await webhook
Attach RCA report
Step 01
An incident or alert triggers the external workflow.
Step 02
n8n, ServiceNow, or another tool starts the RCA request.
Step 03
Pardinus receives the RCA job through the API.
Step 04
The caller waits for completion or polls the job status.
Step 05
The resulting evidence and summary are added to the ticket.
Deterministic-first RCA pipeline
The model does not work from a raw alert alone. Pardinus first builds the incident context, then asks the selected provider to interpret it.
01
Trigger
Launch RCA directly from a Checkmk action on any host or service.
02
Collect
Pardinus gathers status, plugin output, related services, acknowledgements, and downtime context.
03
Structure
The platform assembles the collected signals into a structured incident context.
04
Synthesize
The selected model receives that package and returns a structured response.
05
Report
Teams receive a likely cause, confidence score, and recommended next actions.
Why Pardinus
Pardinus is built for teams that want faster operations without giving up control or deployment flexibility.
Native to Checkmk
Works from the menus, actions, and views teams already use. No separate console and no context switch.
Deterministic Before AI
Pardinus prepares incident context before any model is asked to explain what happened.
Bring Your Own Model
Run with a local LLM, an OpenAI-compatible endpoint, Azure OpenAI, or a customer-managed provider that fits your governance model.
Auditable by Design
Every RCA can show what was used, how confident the result is, and where context is incomplete.
Enterprise guardrails
Pardinus includes practical controls for governance, private deployment, and safer adoption.
Sensitive-value redaction
Sensitive values such as IP addresses, secrets, and tokens can be removed before data leaves the customer environment.
Provider validation
Configured providers can be validated before RCA runs are accepted.
Deterministic fallback
If generation fails, the platform can still return the evidence package so operations can keep moving.
Confidence thresholding
Low-confidence output is flagged for operator review instead of being presented as certainty.
Without protection
Sensitive values remain visible in the payload before generation.
With protection
Sensitive values are masked before data leaves the controlled environment.
Talk to your infrastructure
Pardinus Chat routes natural-language requests to the right capability without exposing teams to API syntax or runbook complexity.
Task
Acts through the Checkmk REST API on the user's behalf and asks for confirmation before destructive operations.
"Create a downtime for all hosts in Lisboa for the next 2 hours."
Report
Reviews system state, service health, and metrics to return a clear operational answer in plain language.
"How are my Linux machines performing today?"
Wiki
Explains Checkmk concepts and configuration topics with session-aware guidance that helps teams onboard faster.
"What is a host tag group?"
Where it creates value
Pardinus fits teams that need faster response, better knowledge capture, and cleaner managed-service delivery.
On-call triage
Critical alerts can trigger RCA so the first responder starts with a clearer incident picture instead of raw plugin output.
Tier-1 escalation
Service-desk teams can attach an RCA report to tickets before handing incidents to L2 or L3.
Recurring incident clustering
Teams can spot repeated patterns across past RCAs and turn recurring failures into known operational issues.
Compliance and post-mortems
Structured output supports ITIL reviews, post-incident records, and regulated operating environments.
MSP multi-tenancy
Per-customer providers and isolated execution make the suite suitable for managed services and white-label offerings.
Living runbooks
Resolved RCA results become reusable operational knowledge that newer team members can search and apply.
Ready to bring Pardinus into your Checkmk operations?
Start with a guided pilot and see how Pardinus can reduce triage time, improve incident clarity, and give teams faster answers.