Scroll to navigation

biolatpcts(8) System Manager's Manual biolatpcts(8)


biolatpcts - Monitor IO latency distribution of a block device.


biolatpcts [-h] [-i INTERVAL] [-w which] [-p PCT,...] [-j] [-v] DEV


biolatpcts traces block device I/O (disk I/O) of the specified device, and calculates and prints the completion latency distribution percentiles per IO type periodically. Example:

# biolatpcts /dev/nvme0n1
nvme0n1 p1 p5 p10 p16 p25 p50 p75 p84 p90 p95 p99 p100
read 95us 175us 305us 515us 895us 985us 995us 1.5ms 2.5ms 3.5ms 4.5ms 10ms
write 5us 5us 5us 15us 25us 135us 765us 855us 885us 895us 965us 1.5ms
discard 5us 5us 5us 5us 135us 145us 165us 205us 385us 875us 1.5ms 2.5ms
flush 5us 5us 5us 5us 5us 5us 5us 5us 5us 1.5ms 4.5ms 5.5ms

biolatpcts prints a number of pre-set latency percentiles in tabular form every three seconds. The interval can be changed with the -i option.

The latency is measured between issue to the device and completion. The starting point can be changed with the -w option.

Any number of percentiles can be specified using the -p option. The input percentile string is used verbatim in the output to ease machine consumption.

-j option enables json output. The result for each interval is printed on a single line.

This tool works by tracing blk_account_io_done() with kprobe and bucketing the completion latencies into percpu arrays. It may need updating to match the changes to the function.

Since this uses BPF, only the root user can use this tool.




-h Print usage message.

Report interval. (default: 3)
Which latency to measure. (default: on-device)
Percentiles to calculate. (default: 1,5,10,16,25,50,75,84,90,95,99,100)
Output in json.
Enable debug output.
Target block device. /dev/DEVNAME, DEVNAME or MAJ:MIN.


# biolatpcts -i 1 sda
# biolatpcts -p 99,99.9,99.99,99.999 -j -i 1 /dev/nvme0n1


This traces kernel functions and maintains in-kernel per-cpu latency buckets, which are asynchronously copied to user-space. This method is very efficient, and the overhead for most storage I/O rates should be negligible. If you have an extremely high IOPS storage device, test and quantify the overhead before use.


This is from bcc.

Also look in the bcc distribution for a companion _examples.txt file containing example usage, output, and commentary for this tool.




Unstable - in development.


Tejun Heo


biolatency(8), biosnoop(8)

2020-04-17 USER COMMANDS