github.com/NVIDIA/aistore@v1.3.23-0.20240517131212-7df6609be51d/docs/cli/performance.md (about) 1 --- 2 layout: post 3 title: PERFORMANCE 4 permalink: /docs/cli/performance 5 redirect_from: 6 - /cli/performance.md/ 7 - /docs/cli/performance.md/ 8 --- 9 10 `ais performance` or (same) `ais show performance` command supports the following 5 (five) subcommands: 11 12 ```console 13 $ ais performance <TAB-TAB> 14 counters throughput latency capacity disk 15 ``` 16 17 ## `ais show performance latency` 18 19 Example usage: 20 21 ```console 22 $ ais show performance latency --refresh 10 23 24 latency ------------------ 13:03:58.329680 25 TARGET GET-COLD(n) GET-COLD-RW(t) GET-COLD(total/avg size) GET(n) GET(t) GET-REDIR(t) GET(total/avg size) 26 t[EkMt8081] 151 2.01s 145.00MiB 983.31KiB 154 2.13s 1.156551ms 154.00MiB 1.00MiB 27 28 latency ------------------ 13:04:08.335764 29 TARGET GET-COLD(n) GET-COLD-RW(t) GET-COLD(total/avg size) GET(n) GET(t) GET-REDIR(t) GET(total/avg size) 30 t[EkMt8081] 189 2.04s 181.00MiB 980.66KiB 190 1.86s 892.015µs 190.00MiB 1.00MiB 31 ``` 32 33 Notice naming conventions: 34 35 * (n) - counter (total number of operations of a given kind) 36 * (t) - time (latency of the operation) 37 38 Other notable semantics includes: 39 40 | metric | comment | 41 | ------ | ------- | 42 | `GET-COLD-RW(t)` | denotes (remote read, local write) latency, which is a _part_ of the total latency _not_ including the time it takes to transmit requested payload to user | 43 | `GET(t)` | GET latency (for cold GETs includes the above) | 44 | `GET-REDIR(t)` | time that passes between ais gateway _redirecting_ GET operation to specific target, and this target _starting_ to handle the request | 45 46 ## `ais show performance counters` 47 48 ```console 49 $ ais show performance counters --help 50 NAME: 51 ais show performance counters - show (GET, PUT, DELETE, RENAME, EVICT, APPEND) object counts, as well as: 52 - numbers of list-objects requests; 53 - (GET, PUT, etc.) cumulative and average sizes; 54 - associated error counters, if any, and more. 55 56 USAGE: 57 ais show performance counters [command options] [TARGET_ID] 58 59 OPTIONS: 60 --refresh value interval for continuous monitoring; 61 valid time units: ns, us (or µs), ms, s (default), m, h 62 --count value used together with '--refresh' to limit the number of generated reports, e.g.: 63 '--refresh 10 --count 5' - run 5 times with 10s interval (default: 0) 64 --no-headers, -H display tables without headers 65 --regex value regular expression select table columns (case-insensitive), e.g.: 66 --regex "put|err" - show PUT (count), PUT (total size), and all supported error counters; 67 --regex "[a-z]" - show all supported metrics, including those that have zero values across all nodes; 68 --regex "(GET-COLD$|VERSION-CHANGE$)" - show the number of cold GETs and object version changes (updates) 69 --units value show statistics and/or parse command-line specified sizes using one of the following _units of measurement_: 70 iec - IEC format, e.g.: KiB, MiB, GiB (default) 71 si - SI (metric) format, e.g.: KB, MB, GB 72 raw - do not convert to (or from) human-readable format 73 --average-size show average GET, PUT, etc. request size 74 ``` 75 76 ## `ais show performance disk` 77 78 ```console 79 $ ais show performance disk --help 80 NAME: 81 ais show performance disk - show disk utilization and read/write statistics 82 83 USAGE: 84 ais show performance disk [command options] [TARGET_ID] 85 86 OPTIONS: 87 --refresh value interval for continuous monitoring; 88 valid time units: ns, us (or µs), ms, s (default), m, h 89 --count value used together with '--refresh' to limit the number of generated reports, e.g.: 90 '--refresh 10 --count 5' - run 5 times with 10s interval (default: 0) 91 --no-headers, -H display tables without headers 92 --units value show statistics and/or parse command-line specified sizes using one of the following _units of measurement_: 93 iec - IEC format, e.g.: KiB, MiB, GiB (default) 94 si - SI (metric) format, e.g.: KB, MB, GB 95 raw - do not convert to (or from) human-readable format 96 --regex value regular expression select table columns (case-insensitive), e.g.: 97 --regex "put|err" - show PUT (count), PUT (total size), and all supported error counters; 98 --regex "[a-z]" - show all supported metrics, including those that have zero values across all nodes; 99 --regex "(GET-COLD$|VERSION-CHANGE$)" - show the number of cold GETs and object version changes (updates) 100 --summary tally up target disks to show per-target read/write summary stats and average utilizations 101 ```