I have an application front-ended by Apache 1.3, and the application cannot
handle more than roughly 200 concurrent requests. We throttled back Apache by
setting MaxClients to 190 and that seems to regulate the load and prevent the
application server from crashing. The problem is, if the number of requests to
Apache exceeds 190 Apache queues those additional requests and users end up
seeing very long response times. Instead of having our users watch the
hourglass we’d like Apache to immediately return “server too busy”
responses (HTTP 503?) as soon as it reaches MaxClients. How can we do this? |
Attachment:
smime.p7s
Description: S/MIME cryptographic signature