Breaker Types

avg

Trips when the arithmetic mean of values in a window exceeds a threshold.

• Monitoring general response time trends.
• Tracking resource utilization where occasional spikes are expected.

consecutive failures

Trips after a specific number of failures occur sequentially.

• Critical external API dependencies.
• Database connection health checks.

count

Trips when the total number of events in a window exceeds a threshold.

• Volume-based rate limiting or DDoS protection.
• Detecting sudden traffic surges.

delta

Trips when the current value diverges significantly from the recent trend.

• Early warning for rapid service degradation.
• Catching spikes where a fixed threshold is too high to be useful.

error rate

Trips when the ratio of errors to total requests exceeds a threshold.

• Detecting degraded service quality in variable-traffic environments.
• Catching partial outages where a service is functional but unstable.

max

Trips when a single value in the window exceeds a ceiling threshold.

• Critical safety limits (e.g., Memory usage at 95%).
• Catching extreme latency outliers.

min

Trips when a single value in the window falls below a floor threshold.

• Detecting zero-traffic outages on critical paths.
• Monitoring baseline performance for heartbeats.

p95

Trips when the 95th percentile (the "slowest 5%") exceeds a threshold.

• Monitoring user-facing latency where worst-case experience is the priority.
• Detecting performance issues affecting specific request subsets.

percentile

Trips when a user-defined percentile of a metric exceeds a threshold.

• Monitoring high-precision tail latency (e.g., p99) for critical services.
• Tracking median performance (p50) to focus on typical user experience.

ratio

Trips when the relationship between two different metrics is imbalanced.

• Capacity planning (Requests per Worker).
• Cost efficiency (Operations per Dollar).

slope

Trips when the long-term trajectory of a metric indicates a steady crawl in a bad direction.

• Detecting slow memory leaks.
• Identifying gradual database performance degradation.

stddev

Trips when the inconsistency of a metric exceeds a threshold.

• Detecting unstable performance despite a healthy average.
• Catching services oscillating between fast and slow states.

sum

Trips when the cumulative total of all values in a window exceeds a budget.

• Monitoring aggregate resource costs (e.g., bytes sent, API credits).
• Budget-based limits for expensive operations.