Alerting SLOs
When humans get paged, and what 'healthy' means numerically — SLOs as the trigger for alerts.
Why we need this / what value this brings
Without a defined threshold, 'is this bad enough to page someone' is a judgment call made under pressure during an incident.
When to use this
Once there's a metric worth being paged about — an SLO gives that page a clear, agreed-upon trigger.
How to use or implement this
Define SLOs against user-facing outcomes (booking success rate), and alert on the SLO being at risk, not on every minor metric blip.
Empty folder — drop notes, links, and findings here as you research.