On-Call Health is an open-source tool designed to catch exhaustion before it burns out engineers by looking for signs of overload in on-call teams. Its main purpose is to provide objective signals that help spot unsustainable workloads early and prevent burnout.

The tool connects to various signals from incident management tools like Rootly or PagerDuty for incident data, Linear for ticket workload, GitHub for after-hours signals, and Slack for communication patterns and context. It also collects sentiment through short, low-friction surveys in Slack, allowing responders to share how they're doing without stigma.

On-Call Health computes individual risk scores from ingested data, categorizing them from 0-24 for maintaining balance, 25-49 for monitoring risk, 50-74 for early intervention, and 75-100 for immediate action. AI analysis identifies what changed and what's driving it to enable informed decisions for protecting engineers.

The system works by pulling data from integrated tools, combining it with self-reported check-ins, and tracking against personal and team baselines to measure trends over time fairly, without relying on fixed thresholds or comparing people.

Benefits include the ability to spot trend shifts early and intervene with small fixes like rebalancing rotations, adding automation, pausing non-urgent work, or staffing up. This helps prevent burnout from becoming reality and makes on-call health actionable.

Target users are on-call engineers and teams, with integrations for Rootly, PagerDuty, GitHub, Linear, Jira, and Slack. The tool is open source under the Apache License 2.0, and AI summaries help stakeholders align faster during weekly incident reviews.