-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 13/11/2014 4:11 p.m., Alberto Klocker wrote: > Looking at the squidclient mgr:info command output I was wondering > what the difference between these two entries are? > > Cache information for squid: Hits as % of all requests: 5min: > 0.7%, 60min: 0.3%z Hits as % of bytes sent: 5min: 51.3%, > 60min: 25.5% > > > I can guess the first one means all requests but I'm stumped on the > wording of the second entry. HTTP messages consist of a variable amount of bytes each. A HIT on a 1KB object saves less bandwidth than a HIT on a 1MB object. The only (relatively) predictable thing based on message count is consumed CPU and network sockets. "% of all requests" is counting by messages. - higher values means less CPU consumed/wasted upstream (server latency), less contention for TCP sockets (HTTP pipeline congestion). "% of bytes sent" is counting by bandwidth used. - higher values means less bandwidth consumed/wasted upstream (TCP congestion). So ... in the past 60min 0.3% of your clients were responded to using the cache, which reduced your upstream bandwidth usage/cost by 25.5%. Amos -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (MingW32) iQEcBAEBAgAGBQJUZCouAAoJELJo5wb/XPRjzJIIAM499MGEXNvMGI60Vul/jShC 2eP5W7O/6rDygQQBkSpI/h9ph0xcgRc4eG+F/jbvUIY828fYOkIZirAOknr8NxPD hV3xB3nZw859SyOfB6+Iz0X2Fp7VWIZZiRg27XxGuWcWsLGF7qmA7Kjh6S0jy21c mCSZYKU5Rgx6FMNNBac7NxztogD0ZD2z2JQwJsSK1JI4JFkw2eImu+NTPPU6aQUG pdiRBGTzATTL0lDXOBaM96s7o3XwkKift4OzZgetBoG+b3LQ2XX4Iw9caBnK/Fr8 MS2nLsvISUI+XrvpqFTYFdTlRVOEYjzfOGPzOwxgNdl86TlzTMh1ExOTmWd92RE= =pN5q -----END PGP SIGNATURE----- _______________________________________________ squid-users mailing list squid-users@xxxxxxxxxxxxxxxxxxxxx http://lists.squid-cache.org/listinfo/squid-users
