Mqttable
Back to docs

Monitor Metrics: runtime context for MQTT decisions

How runtime metrics help separate broker behavior from local machine pressure.

Jun 24, 2026Mqttable EngineeringMqttable Engineering

/monitor/metrics is a product surface in Mqttable. It exists because MQTT debugging needs a place for this job, not because every task should start from MCP. This page focuses on the product workflow. MCP support is a companion layer: use it when automation helps, not as the starting point for every task.

Interactive diagram

Runtime metrics beside broker activity, queue pressure, sockets, and bench context.

Placeholder for the interactive diagram. Replace with the generated artifact when ready.

placeholder

Human job

  • Watch memory, process, queue, socket, and runtime pressure while MQTT work runs.
  • Use metrics to explain Bench or Replay behavior instead of blaming the broker first.
  • Jump to events when a metric needs narrative context.

Automation handoff

  • MCP and CLI are supporting surfaces here, not the whole story.
  • Use MCP reads when the same state needs to be inspected by an agent.
  • If an action changes MQTT state, use schema, plan, confirmation, apply, and verify.

Boundary

Keep this surface focused. If the job belongs to another workflow, move there instead of turning one page into everything.