This bug was tested on both v6 and v7
The gist is that a Pool is not respecting keep-alive timeouts and closes connections once the request queue is empty. For example, if you create a pool and execute a batch of requests, then execute another batch of requests after some timeout less than either the configured keepAliveTimeout option or the response keep-alive directive, the connections are not reused from the first batch.
I've created a test case in PR #4332 demonstrating the issue.
This bug was tested on both v6 and v7
The gist is that a
Poolis not respecting keep-alive timeouts and closes connections once the request queue is empty. For example, if you create a pool and execute a batch of requests, then execute another batch of requests after some timeout less than either the configuredkeepAliveTimeoutoption or the responsekeep-alivedirective, the connections are not reused from the first batch.I've created a test case in PR #4332 demonstrating the issue.