moonchen opened a new pull request, #13272:
URL: https://github.com/apache/trafficserver/pull/13272
These counters feed a per-component CPU model that attributes ssl:accept
and ssl:close User CPU to its causal workload drivers instead of folding
it into a fixed intercept.
The dominant cost of a TLS accept is the asymmetric signature, which only
a full handshake performs -- a resumed handshake skips it. Splitting the
existing proxy.process.ssl.total_success_handshake_count_in into _in_full
and _in_resumed, and recording the server private-key type (RSA being far
heavier than ECDSA), exposes that cost as a workload count.
New counters:
* proxy.process.ssl.total_success_handshake_count_in_full
* proxy.process.ssl.total_success_handshake_count_in_resumed
* proxy.process.ssl.handshake_sign_rsa / _ecdsa / _other
* proxy.process.ssl.connections_closed -- one per SSL_free, the driver
behind ssl:close where cert-chain teardown dominates.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
