Handling Overload¶
Beyer, B., Jones, C., Petoff, J., & Murphy, N. R. (2016). Handling Overload: How Google Runs Production Systems. Site Reliability Engineering: How Google Runs Production Systems.
Cited by¶
1 citation across 1 artifact.
Each citation links to the sentence it supports in the citing article.
Primes¶
- Input Pressure
- On the driver side, the rate can be measured as a first-class metric (separate from cumulative quantity), forecast as a distribution rather than a current value, capped at the source (effluent standards, rate limiting, admission control, arrival smoothing), buffered over time to convert a peak-rate problem into an average-rate problem (reservoirs, queues, stockpiles), or shed gracefully when capacity is exceeded (overflow weirs, drop policies, triage).
This sourceDescribes the driver-side overload toolkit — admission control, client-side throttling, load shedding, and graceful degradation — for capping and shedding a sustained request rate.
- On the driver side, the rate can be measured as a first-class metric (separate from cumulative quantity), forecast as a distribution rather than a current value, capped at the source (effluent standards, rate limiting, admission control, arrival smoothing), buffered over time to convert a peak-rate problem into an average-rate problem (reservoirs, queues, stockpiles), or shed gracefully when capacity is exceeded (overflow weirs, drop policies, triage).
Verification¶
This reference passed the adversarial substantiation pipeline: it was checked to exist and to support the claim it is attached to. See how references were verified.
Registry ID ref:7ee090f8b4e3 · see in the full table