sentinel限流 在高并发调用下。可能会限流不准

我看源码里是否要限流。如下代码判断：

StatisticSlot.entry方法：
fireEntry(context, resourceWrapper, node, count, prioritized, args);

// Request passed, add thread count and pass count.
node.increaseThreadNum();
node.addPassRequest(count);

这里是先进行了 限流判断。再决定是否累加qps数。

DefaultController.canPass方法：
int curCount = avgUsedTokens(node); //获取当前的流量统计
if (curCount + acquireCount > count) {
if (prioritized && grade == RuleConstant.FLOW_GRADE_QPS) {
long currentTime;
long waitInMs;
currentTime = TimeUtil.currentTimeMillis();
waitInMs = node.tryOccupyNext(currentTime, acquireCount, count);
if (waitInMs < OccupyTimeoutProperty.getOccupyTimeout()) {
node.addWaitingRequest(currentTime + waitInMs, acquireCount);
node.addOccupiedPass(acquireCount);
sleep(waitInMs);

// PriorityWaitException indicates that the request will pass after waiting for {@link @waitInMs}.
throw new PriorityWaitException(waitInMs);
}
}
return false;
}
return true;

假设限流设置1000个。假如进来2000个请求 同时执行到 canpass方法的 if (curCount + acquireCount > count)。那么这两千个请求 都能通过。

Start with StatisticSlot.entry and DefaultController.canPass, focusing on the ordering of the limit check and the subsequent request-count updates. Reproduce the reported case with a limit of 1000 and 2000 concurrent requests, then inspect the existing flow-control tests. Done means the concurrency case enforces the configured limit and has regression coverage.

Written by the indexing model from the issue text.