max_clients limit reached, alerts reset

I occasionally see this in the log for one of my metrics. What I noticed as a side affect that it never clears the alert if it was previously in a Warning state, and then goes into normal area, the normal alert doesn't fire, maybe it's the reset alerts?

Is reset alerts related to the max_clients limit reached? Is there a way to increase the size?

`
graphite_beacon stderr | [D 160524 16:32:00 simple_httpclient:125] max_clients limit reached, request queued. 10 active, 1 queued requests.
[I 160524 16:32:00 core:117] Reset alerts
`

Start by tracing the `simple_httpclient:125` and `core:117` log points shown in the report, focusing on the relationship between queued requests, the max_clients limit, and alert resets. Reproduce or inspect the warning-to-normal transition and determine whether alerts clear and fire correctly after the reset; document the relevant configuration or behavior.

Written by the indexing model from the issue text.