# From Incident Counting to SLIs: How DigitalOcean Rethought Availability

DevFeed: [From Incident Counting to SLIs: How DigitalOcean Rethought Availability](<https://devfeed.tech/articles/from-incident-counting-to-slis-how-digitalocean-rethought-availability-19946.md>)

Original publisher: [Read original article](<https://www.digitalocean.com/blog/sli-based-availability-framework>)

Author: Miguel Carrera

Published: 2026-04-23T09:15:00Z

Content type: article

Language: en

Sources: [DigitalOcean](<https://devfeed.tech/sources/digitalocean.md>)

Topics: [Availability](<https://devfeed.tech/topics/availability.md>), [Digital Ocean](<https://devfeed.tech/topics/digital-ocean.md>), [incident](<https://devfeed.tech/topics/incident.md>), [Framework](<https://devfeed.tech/topics/framework.md>), [Prometheus](<https://devfeed.tech/topics/prometheus.md>)

Tags: [availability](<https://devfeed.tech/tags/availability.md>), [digitalocean](<https://devfeed.tech/tags/digitalocean.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [framework](<https://devfeed.tech/tags/framework.md>), [incident](<https://devfeed.tech/tags/incident.md>), [metric](<https://devfeed.tech/tags/metric.md>), [platform](<https://devfeed.tech/tags/platform.md>), [prometheus](<https://devfeed.tech/tags/prometheus.md>)

## AI overview

DigitalOcean replaced incident-based availability tracking with an SLI-driven framework covering the control plane, data plane, and error budgets. The article explains that incident counting did not reflect customer experience and describes the newer measurement approach.

## Source excerpt

Our journey to truly understand our customer experience began with a hard look at our internal availability numbers at the start of 2025. We saw something uncomfortable: the numbers didn't match our customers' reality. Our monthly availability oscillated between 99.5% and 99.9%. Those peaks and valleys depended more on whether we declared a high-severity incident that month than on how the platform was actually performing. Customers were still experiencing issues and opening escalations, but the metric didn't reflect customer availability. The previous internal measurement served us well in our early days, but its limitations became evident as DigitalOcean expanded. Our incident-based approach treated any declared incident as a total outage and anything below the severity threshold as invisible. This created a structural trap: we couldn't expand coverage to include lower-severity issues without artificially destroying our availability number, because the formula would count every minute of a partial degradation as a full platform outage. The chart above shows monthly platform availability using both methodologies over the same time period. The incident-based (old) swings between roughly 99.5% and 99.9% month to month. The SLI-based metric (new) holds consistently at 99.95% or above. The old metric was measuring noise, while the new metric measures actual availability signals. This isn't a problem unique to DigitalOcean. Any platform that measures availability by counting incident minutes against total calendar time will eventually hit the same wall. The incident-based metric was both too generous and too punitive, depending on where the line was drawn. Thisarticle walks through the operational framework we built to replace that system, the architectural decision to split the measurement into Control Plane and Data Plane, the two different SLI methodologies we use for each plane, the Prometheus recording rules and multi window alerting that make it operational, error