• Choose a category
• All
• Classic-Fiction

average_select_fd_period

The mean number of seconds required to scan a file descriptor in the main I/O loop.

median_select_fds

The five-minute median number of ready file descriptors each time Squid calls select( ) or poll( ) (the median of the select( )/poll( ) return value). Unfortunately, this value is almost always zero because Squid's functions for calculating the median don't work very well with the select_fds histogram, in which 0 and 1 are the most common values.

syscalls.selects

The five-minute mean rate of calls to select( )/poll( ). If Squid is using poll( ) on your system, the variable is called syscalls.polls. This value may be a little larger than select_loops, because the latter only includes calls in the main I/O loop.

syscalls.disk.opens

The five-minute mean rate of open( ) calls for disk files.

syscalls.disk.closes

The five-minute mean rate of close( ) calls for disk files.

syscalls.disk.reads

The five-minute mean rate of read( ) calls for disk files.

syscalls.disk.writes

The five-minute mean rate of write( ) calls for disk files.

syscalls.disk.seeks

The five-minute mean rate of seek( ) calls for disk files. Probably zero unless you are using aufs, which always calls seek( ) before reading.

syscalls.disk.unlinks

The five-minute mean rate of unlink( ) (or, in some cases, truncate( )) calls for disk files.

syscalls.sock.accepts

The five-minute mean rate of accept( ) calls for network sockets.

syscalls.sock.sockets

The five- minute mean rate of socket( ) calls for network sockets.

syscalls.sock.connects

The five-minute mean rate of connect( ) calls for network sockets.

syscalls.sock.binds

The five-minute mean rate of bind( ) calls for network sockets.

syscalls.sock.closes

The five-minute mean rate of close( ) calls for network sockets.

syscalls.sock.reads

The five-minute mean rate of read( ) calls for network sockets.

syscalls.sock.writes

The five-minute mean rate of write( ) calls for network sockets.

syscalls.sock.recvfroms

The five-minute mean rate of recvfrom( ) calls for network sockets. Used for UDP-based protocols, such as DNS, ICP, HTCP, and some interprocess communication.

syscalls.sock.sendtos

The five-minute mean rate of sendto( ) calls for network sockets. Used for UDP-based protocols, such as DNS, ICP, HTCP, and some interprocess communication.

60min: 60 Minute Average of Counters

This page shows a 60-minute average of the data in the Traffic and Resource Counters page. The descriptions are identical to those for the 5 Minute Average of Counters page, except the measurements are taken over one hour.

utilization: Cache Utilization

This page displays averages of the counters (see Traffic and Resource Counters and 5 Minute Average of Counters) over various time spans. The same values are reported for 5-minute, 15-minute, 1-hour, 8-hour, 1-day, and 3-day intervals.

This page, with a poorly chosen name, exists primarily so that developers can take a quick snapshot of statistics for testing purposes.

histograms: Full Histogram Counts

This page displays the current histogram values (since Squid was started) for a number of measurements:

client_http.all_svc_time

client_http.miss_svc_time

client_http.nm_svc_time

client_http.nh_svc_time

client_http.hit_svc_time

icp.query_svc_time

icp.reply_svc_time

dns.svc_time

select_fds_hist

These are the same measurements described in Section 14.2.1.33, except that here Squid gives the full histogram, instead of the mean or median.

Depending on the type of histogram, you may see two or three columns. The first column is the bin number and lower bound on the bin value. The second column is the number of counts for that bin. The optional third column is the number of counts divided by the "size" of the bin. The last column is probably only interesting for log-based histograms, in which the bin size isn't constant.

active_requests: Client-Side Active Requests

This page shows a list of currently active