mirror of https://git.kore.io/kore.git
Before http_request_limit just constrained the number of HTTP requests we'd deal with in a single http_process_requests() call. But it should really mean how many maximum HTTP requests are allowed to be alive in the worker process before we start sending 503s back. While here, drop the lock timeout for a worker to 100ms down from 500ms and do not allow a worker to grab the accept lock if their HTTP request queue is full. This makes things much more pleasant memory wise as the http_request_pool won't just grow over time. |
||
---|---|---|
.. | ||
accesslog.c | ||
auth.c | ||
bsd.c | ||
buf.c | ||
cli.c | ||
config.c | ||
connection.c | ||
domain.c | ||
http.c | ||
jsonrpc.c | ||
keymgr.c | ||
kore.c | ||
linux.c | ||
mem.c | ||
module.c | ||
msg.c | ||
net.c | ||
pgsql.c | ||
pool.c | ||
python.c | ||
runtime.c | ||
tasks.c | ||
timer.c | ||
utils.c | ||
validator.c | ||
websocket.c | ||
worker.c |