From deep MongoDB monitoring and a major upgrade to LLM Observability, to OpsAI integrations inside your IDE and AI-powered workflows across the platform, this release is packed with features that help you investigate, build, and monitor faster.

Here’s everything new this month 👇

🤖 OpsAI

We expanded OpsAI beyond issue investigation, making it available across the entire platform and directly inside your development workflow.

Remote MCP Support

Middleware now provides a Remote MCP Server, allowing developers to access observability data directly from their IDE or AI coding assistants.

With MCP, you can:

  • Query logs, traces, metrics, and alerts without leaving your editor
  • Investigate production issues directly from your development environment
  • Bring live observability context into AI-assisted coding workflows

 

Here is the doc for integration.

Run Kubernetes Commands During Investigation

OpsAI can now execute Kubernetes commands directly from an investigation or chat.

  • One-click execution
  • No need to switch to a terminal
  • Faster debugging and remediation workflows
    Kubernetes Run Command

Create Dashboards from Chat

Simply describe the dashboard you need in the chat, and OpsAI will:

  • Generate the dashboard
  • Configure widgets automatically
  • Return a link to the newly created dashboard

Vercel Support

OpsAI now supports:

  • Vercel logs
  • Fatal errors

These issues appear directly in the OpsAI issue list, allowing you to investigate them and generate GitHub pull requests with suggested fixes.

Synthetic Monitoring with OpsAI

OpsAI can now answer questions about Synthetic Monitoring and create synthetic tests directly from chat.

Examples include:

  • Creating uptime checks
  • Reviewing test results
  • Troubleshooting failed synthetic tests

Ask OpsAI from Anywhere

Every major screen across Middleware now includes an Ask OpsAI button.

Ask questions about the data you’re viewing, or simply ask OpsAI to:

  • Create dashboards
  • Create alerts
  • Explain anomalies
  • Analyze metrics, logs, or traces

without leaving your current screen.

RUM Data Support

OpsAI now understands Real User Monitoring (RUM) data.

Start a new chat to:

  • Analyze frontend performance
  • Investigate user sessions
  • Understand browser errors
  • Ask questions about user behavior

Navigation Updates

To simplify navigation:

  • Removed Error & OpsAI from the left navigation menu
  • Updated the Welcome Dashboard with additional filters for quicker access to investigations

🍃 MongoDB Database Monitoring

We’ve introduced several new screens to help teams monitor database performance in greater detail.

New Monitoring Views

  • Query Metrics
  • Blocking Queries
  • Query Samples

Along with new MongoDB performance metrics to monitor overall database health and workload characteristics.

📜 Log Monitoring

Filter from RAW JSON

You can now apply filters directly from the RAW JSON view using parsed attributes, making log exploration much faster.

Log monitoring filter

Log Anomaly Detection

Detect anomalies such as:

  • Count spikes
  • Behavior drift
  • Pattern inactivity
  • Other unexpected log behavior
    Log annomaly

Each anomaly includes a dedicated detail view showing:

  • Related log properties
  • Matching logs
  • AI-generated analysis and insights

🧠 LLM Observability – Major Upgrade

We’ve significantly expanded our LLM Observability capabilities.

Richer LLM Traces

LLM trace details now include:

  • Total tokens per request
  • Cost
  • Model information
  • Span details
  • Stack traces
  • Chat input
  • Related logs

Giving complete visibility into every LLM request.

Playground

Experiment with different LLM configurations directly inside Middleware.

You can:

  • Choose different models
  • Define output schemas
  • Add tools
  • Test prompts interactively

before deploying them into production.

LLM as a Judge (Evaluation)

Evaluate prompts automatically using another language model.

Configure:

  • Evaluation model
  • Acceptance criteria
  • Scoring rules
  • Evaluation scope

OpsAI will judge each prompt and measure its quality against your configured rules.

LLM Dashboard

A brand-new dashboard provides comprehensive visibility into your AI workloads.

Monitor over 50+ charts, including:

  • Token usage
  • Cost
  • Latency
  • Error rate
  • LLM calls
  • Evaluations
  • Tool usage
  • Request duration

🎮 GPU Monitoring

Monitor GPU infrastructure with detailed device-level insights.

Track metrics including:

  • GPU utilization
  • Occupancy
  • Memory usage
  • Temperature
  • Power consumption
  • Bandwidth
  • Running processes

Ideal for AI and ML workloads running on GPU-enabled infrastructure.

🐧 Linux Instrumentation

The Linux Instrumentation experience is now more streamlined.

You can configure:

  • Detected databases
  • Integrations

directly from the Linux Instrumentation screen without navigating to the Integrations page.

🔌 Kafka Integration

Kafka monitoring now supports Kubernetes deployments.

Monitor Kafka clusters running inside Kubernetes with native integration support.

☁️ S3 Routing

You can now route observability data directly to your own AWS S3 bucket.

Supported data types include:

  • Metrics
  • Logs
  • Traces

A new Settings → S3 Routing page lets you configure:

  • AWS Account ID
  • Bucket name
  • AWS Region

making it easy to archive or export observability data to customer-managed storage.