SentinelOps
Grounded SRE Answers from MongoDB Atlas Vector Search
AI-powered incident diagnosis combining MongoDB Atlas Vector Search, Gemini 2.5, and MCP for factually-grounded responses backed by your runbook library.
Semantic Vector Search
768-dimensional embeddings enable semantic matching of incidents to relevant runbooks with cosine similarity scoring.
AI Memory Synthesis
Gemini automatically generates personalized user profiles by analyzing conversation history and context patterns.
Transparent Grounding
Every response shows which runbooks were used with exact similarity scores—no hallucinations, just facts.
MCP Tool Orchestra
25 MongoDB MCP tools orchestrated by Gemini for database reads, writes, aggregations, and vector search operations.
Powered By
🚨 Active Infrastructure Incidents (Dynatrace Simulation)
Server Latency Spike — us-central1 CPU Usage 98.4%
Dynatrace detected a latency regression of +1200ms in core microservice.
MongoDB Connection Timeout — auth-service
Simulated connection lock successfully debugged and resolved by agent.
📈 Live System Performance Metrics DEMO DATA
Awaiting observability alert webhook to execute autonomous root cause analysis and GitLab merge requests.
🔔 Webhook Alert Tester
Test the observability webhook endpoint with a sample alert payload.
Loading...
💬 SentinelOps Diagnostic Chat
🦊 GitLab Merge Request Hotfix (AI-Generated) SIMULATION
# AI SRE Hotfix Git Diff
# Awaiting active incident diagnosis to generate resolution patch...
🧠 MongoDB Atlas Memory Brain
SentinelOps uses a unified MongoDB Atlas cluster to store long-term contextual memories, incident histories, and dense vector embeddings side-by-side.
vector_index (Similarity: Cosine, Dimensions: 768)
{
"title": "SRE Incident Response Protocol",
"content": "SRE Incident Response Protocol - Phase 4. When an observability tool (such as Dynatrace) triggers a webhook alert indicating high server latency or an active CPU spike, the SRE Incident Agent must immediately execute log analysis, identify the breaking commit in the codebase, write a hotfix, and automatically open a GitLab Merge Request containing the code resolution.",
"embedding": [0.0827, -0.0152, 0.1042, 0.0516, -0.1192, 0.0234, ... (768 dimensions)]
}
🛡️ System Grounding Runbook Ingester
Upload, partition, and embed enterprise SRE runbooks, deployment guides, or troubleshooting wikis directly into your MongoDB Atlas Vector Search Database.