The Complete Platform for OpenTelemetry Fleet Management

The Fleet Overview gives you real-time visibility into every connected OpenTelemetry Collector. Toggle dynamically between an interactive honeycomb grid and a rich, sortable data table.

Supervisors maintain a persistent WebSocket connection to the Management Server over OpAMP (port 4320), reporting granular process metrics and heartbeat status every 5 seconds.

Execute lifecycle commands directly from the dashboard. The server translates actions into signed OpAMP command frames that the edge Supervisor executes locally with zero network friction.

The next-generation pipeline visualizer automatically groups active receivers, chained processors, exporters, and connectors into clean, centered execution stages without messy diagonal wires.

Clicking any node in the topology canvas opens an anchored slide-over card displaying the component's exact YAML sub-block extracted directly from the active configuration.

Spin up isolated OTel sidecar pipelines alongside your primary collector. Sidecar binaries run on dedicated ports (14317/14318) so experimental AI or heavy compliance processing never impacts core telemetry.

CollectorCtrl compiles the final Effective YAML for every compute node using an enterprise 4-tier inheritance engine, serving configs over OpAMP in under 2ms from PostgreSQL.

Deploy configuration updates to a canary ring first. Monitor real-time health metrics, then promote to 100% or click rollback to revert the entire fleet in milliseconds.

If an engineer edits a collector configuration file directly on an edge server, the Supervisor detects the hash divergence within seconds and automatically restores the authorized policy.

Every collector binary upgrade is governed by a state machine (offered ➔ verifying ➔ healthy | rolled_back | failed). If an agent fails to report healthy within 3 minutes, the previous binary is re-offered automatically.

CollectorCtrl performs magic-byte binary header inspection on every package upload, verifying executable format (PE, ELF, Mach-O) and matching target CPU architectures.

The built-in OpenTelemetry Collector Builder (OCB) lets you compile minimal, secure collector binaries containing only the components your infrastructure requires.

A persistent, fleet-wide AI assistant available on every page (Ctrl + /). It answers questions in plain English using your live fleet data through a secure toolbelt, shows an Agent Reasoning trace for every answer, and proposes changes as inline approval cards — nothing executes without a human decision.

CollectorCtrl provides a native MCP endpoint (/api/mcp) over JSON-RPC 2.0. Query agent health, telemetry volume, package versions, and audit logs using natural language from any MCP client.

Integrated directly into the Monaco YAML editors. Click ✨ AI Assistant to explain unfamiliar pipeline blocks, diagnose validation errors, or suggest sampling rules.

When AI agents propose configuration updates or policy modifications, CollectorCtrl stages them in a dedicated review queue requiring authorized human approval before application.

Deploy non-invasive edge taps that sample every severity at one rate, then rank the recurring log patterns and span routes that make up most of your telemetry. Identify fragmented logs and collector noise, and generate tested OpenTelemetry rules to fix them at the edge, without the overhead of permanent raw storage.

The Telemetry Governor writes structured, replayable telemetry to Amazon S3 (or MinIO) in Hive-partitioned gzip JSON Lines (.jsonl.gz) for multi-year compliance and ad-hoc query analysis.

The Supervisor automatically discovers running services on the node and injects OpenTelemetry agents into application processes without requiring source code modifications or recompilation.

Built from the ground up for enterprise compliance. Integrate with your corporate identity provider, enforce least-privilege RBAC, and stream every administrative action to your SIEM.