Glance

Glance is an expectation monitoring platform for automated work and AI-run operations.

It answers the operational question that uptime dashboards and issue trackers usually miss: did the thing that was supposed to happen actually happen? Glance lets teams define expectations for jobs, agents, queues, and business signals, then watches for silent failures like missed check-ins, low-output runs, stalled agents, or pipelines that quietly run dry.

What Glance does now:

• Tracks scheduled jobs and external automations through API-backed run reporting
• Computes green/yellow/red status from cadence, failures, duration, and output expectations
• Checks run metadata, such as whether an import produced enough records to count as healthy
• Sends alerts and daily digests while supporting mute/disable/throttle controls
• Connects failures to remediation work items so humans or agents can investigate and fix the problem

The original agent-first work queue is still part of the product, but it has become the response loop: when reality stops matching expectations, Glance opens the work needed to bring the system back into shape.