What Is Cache Stampede Protection?
Cache stampede protection is a set of techniques that stop a crowd of clients from simultaneously recomputing the same expensive value when a popular cache entry disappears. Common approaches let one client recompute while others wait or serve a slightly stale value. The aim is to avoid the thundering-herd load spike a naive cache suffers.
Why it matters
When a hot key expires, every concurrent request misses and tries to regenerate it at once, hammering the backing system exactly when it is busiest. Protection mechanisms like single-flight locking or early recomputation keep that regeneration to one worker. In CI, a stampede on a shared cache backend can stall many runners at the same moment.
Related guides
What Is Cache Warming?Cache warming pre-populates a cache before it is needed so the first real requests hit a ready cache instead…
What Is a Negative Cache?A negative cache stores the fact that a lookup failed or a value is absent, so repeated misses can be answere…
What Is a Remote Cache Backend?A remote cache backend is shared storage where build and test caches live so many machines and CI runs can re…