irtt-rs/docs/irtt-client.md

irtt-client

NAME

irtt-client - IRTT-compatible stream client

SYNOPSIS

irtt-client [OPTIONS] [[LABEL=]TARGET[@hmac=KEY]]...

irtt-client --list-columns

DESCRIPTION

irtt-client probes one or more IRTT-compatible servers over UDP and prints per-probe measurement events as they arrive, plus a final summary for eligible runs.

TARGETS

Every target argument accepts [LABEL=]TARGET[@hmac=KEY]. An optional LABEL= prefix assigns the logical target name used in output; without it, the target string itself is used as the label. Labeled and unlabeled targets can be freely mixed in one argument list. Prefer explicit labels for long-running dynamic target sets: the label is the stable target identity across replacements.

irtt-client host-a:2112
irtt-client eu=host-a:2112 us=host-b:2112
irtt-client host-a:2112 eu=host-b:2112
irtt-client --hmac default-secret eu=host-a:2112@hmac=eu-secret public=host-b:2112@hmac=

Targets without @hmac= inherit the global --hmac value. @hmac=KEY overrides it for that target, and @hmac= explicitly disables HMAC for that target. The modifier uses @, not a shell command separator; quote an argument if its key needs shell quoting.

Target syntax supports these backslash escapes in labels, addresses, and HMAC values: \\ for a literal backslash, \@ for a literal @, \= for a literal =, and \; for a literal semicolon. Escaping @ is necessary in a label or address when it would otherwise begin @hmac=; escaping the first = is how an unlabeled target or a label can contain that character. An = in an HMAC value, including Base64 padding, does not need escaping. Positional target arguments may contain a raw comma; stdin target sets have the additional \, escape described below.

Shells process unquoted backslashes and characters such as semicolons before the client sees them. Single-quote a target when it contains parser escapes, for example 'west\=edge=host\@hmac=literal@hmac=a\;b\\c', so the backslashes reach irtt-client unchanged.

Multiple targets are probed concurrently. By default, a multi-target session uses staggered pacing, spacing active targets across the probe interval; pass --pacing burst to send one probe to every active target back-to-back once per interval instead. Final per-target summaries print in the order targets were supplied on the command line, not the order in which they finish or their labels sort alphabetically.

ADDRESS FAMILY

By default, each target uses the first address returned by the system resolver, whether IPv4 or IPv6. Use -4 / --ipv4 to select IPv4 only, or -6 / --ipv6 to select IPv6 only. Hostnames are resolved when each session connects; these flags do not perform an extra discovery lookup. An explicit IP literal from the other family is rejected during preparation.

Use --dual-stack to compare IPv4 and IPv6 paths as separate measurements:

irtt-client -6 host.example
irtt-client --dual-stack cov=uk-cov1.irtt.lochnair.net

For each hostname, the client discovers which address families the system resolver currently returns. IPv4 results create cov/v4, IPv6 results create cov/v6, and both create both targets, in that order. Without an alias, host.example becomes host.example/v4 and/or host.example/v6. Multiple addresses within one family still produce only one target. Explicit IP literals keep their original label and remain single targets.

Each target has its own session, statistics, and failures. Its original hostname is retained: every new session resolves it again and selects the first address matching its family. Families are discovered only when a declaration is added, so a newly appearing family does not create another target automatically. Discovery failure or a successful lookup returning no addresses rejects the whole target set. A missing family is normal and simply creates no target for that family. Final labels must be unique, including generated /v4 and /v6 labels; choose another alias if one collides with an explicit label.

-4, -6, and --dual-stack are mutually exclusive. Expanded targets use ordinary multi-target pacing and each sends its own probes, so dual-stack can double the number of probes per interval compared with one target per hostname.

DYNAMIC TARGET SETS FROM STANDARD INPUT

--targets-stdin is available only with continuous mode (--duration 0). Positional targets remain an initial desired set, and may be omitted to start empty:

irtt-client --duration 0 --targets-stdin
irtt-client --duration 0 --targets-stdin ams=ams.example:2112

Each non-empty stdin record declares the complete desired target set, rather than a delta. A later record replaces an earlier set: targets absent from it are retired, unchanged target configurations retain their generation, and a changed address or HMAC setting creates a fresh generation. Under transient live-generation backpressure, only the latest unapplied desired set is kept.

The same address-family preparation applies to initial and stdin targets. With --dual-stack, unchanged labels and endpoints retain their discovered families across records, including HMAC-only changes. Removing a declaration and later adding it again performs fresh discovery. New declarations are prepared asynchronously while existing measurements continue. A preparation error stops the stream gracefully without applying a partial set.

Commas frame targets in stdin records; escape a literal comma within one target as \,. Target parsing inside each element is otherwise the same as for positional arguments, including the other escapes above, @hmac=KEY, and @hmac=. Use a persistent producer or keep a pipe/FIFO writer open while the client is being controlled. For example, in one terminal:

mkfifo /tmp/irtt-targets
irtt-client --duration 0 --targets-stdin --hmac default-secret \
  < /tmp/irtt-targets

Then open the FIFO once from another terminal and keep that descriptor open while publishing complete desired sets:

exec 3> /tmp/irtt-targets
printf '%s\n' 'ams=ams.example:2112,sg=sg.example:2112@hmac=sg-secret' >&3
printf '%s\n' 'sg=sg.example:2112@hmac=sg-secret,nyc=nyc.example:2112' >&3
# Publish more sets as needed, then send EOF and stop the client:
exec 3>&-

The exact record [] selects an empty desired set. Empty records are ignored; whitespace is otherwise payload, not trimming. EOF is immediate loss or closure of the control plane and requests a graceful stop. A desired-state revision that was read but has not yet been acknowledged as applied may be discarded when EOF is observed; EOF does not wait for a capacity-rejected revision to become applicable.

The maximum stdin record payload is 64 KiB. Stdin-controlled mode bounds each desired set to 128 targets and retains at most 256 live target generations while replacements drain. The 128-target limit applies after dual-stack expansion: a hostname producing both families consumes two target slots.

FINITE VERSUS CONTINUOUS OPERATION

--duration controls the run length:

• Any nonzero duration (default 10s) is a finite run: probing stops after that long, and a final summary is always printed for table output (see OUTPUT below — CSV, TSV, and JSON Lines never print one).
• --duration 0 is continuous: the client runs until interrupted (Ctrl-C). For table output, a final summary is printed only if the run was interrupted. Stdin-controlled mode is the exception: EOF also requests a graceful stop and prints the final retained summary. An uninterrupted continuous run that exits for another reason does not get one.

See MEASUREMENTS AND MEMORY below for why this distinction also determines what statistics can be computed and how much memory a long run retains.

OUTPUT

--format FORMAT

One of four event-row formats, default table:

• table: human-readable terminal output; the only format that prints a final summary.
• csv: comma-separated output.
• tsv: tab-separated output.
• jsonl: one JSON object per event.

CSV, TSV, and JSON Lines default to all columns for structured export; table defaults to a compact column set and hides echo_sent rows. An explicit --columns selection other than default shows all event rows, including echo_sent; --columns default is equivalent to omitting the option and keeps echo_sent hidden for table output.

--columns COLUMNS / -c COLUMNS

Comma-separated event row columns, or the special values default (the format's default set) or all (every column). List every available column, its meaning, and its aliases with:

irtt-client --list-columns

(--list-columns does not require a target.)

--header auto|always|never

Header policy for table, csv, and tsv output (default auto). JSON Lines never prints a header line.

--verbose

Includes extra fields in table output and final summaries. Has no effect on CSV, TSV, or JSON Lines output either way — those formats show their default all-column set or an explicit --columns selection exactly as given, regardless of --verbose.

Examples

irtt-client <server> --format table
irtt-client <server> --format jsonl
irtt-client <server> --format csv --columns event,seq,remote,effective_rtt_us

A stream containing only effective RTT values in microseconds:

irtt-client <server> --format tsv --columns effective_rtt_us --header never

MEASUREMENT FIELDS

Useful measurement columns:

• raw_rtt_us: client-observed send-to-receive RTT. Always userspace-timed and unaffected by kernel timestamp preference (below).
• adjusted_rtt_us: RTT adjusted for server processing when available. Can be negative when server processing exceeds the measured raw RTT.
• effective_rtt_us: adjusted_rtt_us when available, otherwise raw_rtt_us.
• rd_us / sd_us (selectable with the longer aliases receive_delay_us / send_delay_us, but always emitted under the short names): one-way delay estimates. Can be negative because of clock skew between client and server.
• ipdv_us: inter-packet delay variation.
• server_processing_us: time spent processing the packet at the server.
• kernel_rx_ns: kernel receive timestamp, in Unix nanoseconds, when captured (see below).

Kernel timestamps (Linux)

On Linux, one-way delay estimates prefer a kernel-captured timestamp over userspace timing for this client's own send/receive endpoint, where the platform and socket support it, reducing the effect of scheduling and queuing delay. The peer's reported endpoint gets the same preference when it is the irtt-server applet running on Linux; against another IRTT-compatible server, whatever wall-clock value it reports is used as-is. This is best-effort: an unsupported or implausible kernel timestamp falls back to userspace timing. raw_rtt_us is unaffected either way; adjusted_rtt_us/effective_rtt_us can still shift slightly, since server processing time is derived from the same server receive endpoint.

MEASUREMENTS AND MEMORY

Finite and continuous modes make a different tradeoff between exact statistics and bounded memory.

Finite mode retains exact timing samples (and adjacent-sequence IPDV state) so the final summary can report exact medians. Retained memory therefore grows with the number of probes sent, at a small constant cost per probe (on the order of a few hundred bytes of retained sample state per probe per target, before allocator overhead) — a very long finite run at a short interval is the case to plan capacity for. The client warns on stderr when its own estimate of this retained state crosses roughly 128 MiB, and again more strongly past roughly 512 MiB and 1 GiB; treat those numbers as practical guidance, not a promised RSS ceiling, since allocator and OS overhead sit outside the estimate.

Continuous mode (--duration 0) does not retain a growing sample history: running statistics keep count/min/max/mean/variance only, so exact medians are unavailable. It still keeps a bounded adjacent-sequence IPDV store, capped at 4,096 sequences per target, so nearby replies can still form IPDV pairs. Retained statistics state therefore approaches a fixed per-target bound rather than growing with elapsed run time.

Reply classification (independent of the statistics choice above) is also bounded per target: each of the pending, timed-out, and completed/duplicate sequence stores has its own 4,096-entry limit — not one shared total across the three. A full pending store makes the client fail and drain that target as resource-exhausted rather than wait for capacity to free. Timed-out and completed state evict their oldest entries, so a sufficiently old late reply can still be seen and counted, but no longer has retained send state for measurements such as RTT.

Overall, retained state scales with the number of configured targets and these fixed per-target limits, not with total probes sent in continuous mode. Output is written as events arrive rather than accumulated; the managed event stream feeding it is itself bounded and lossy (16,384 events), so a sufficiently slow consumer can miss events, in which case the client reports that resulting output/statistics may be incomplete.

For capacity planning on constrained devices (for example an OpenWrt router): a finite run's statistics memory is roughly proportional to targets × probes_sent, while a continuous run's statistics memory approaches a fixed multiple of targets regardless of how long it runs.

PEER CLOSE AND EXIT BEHAVIOR

A peer closing a finite run is accepted as a terminal, successful outcome. In continuous mode, an uninterrupted peer closure is treated as an error — the client exits nonzero — so a process supervisor can restart it; an interrupted run (Ctrl-C) never reports peer closure as an error, whichever mode it was in.

With --targets-stdin, peer closure is target-local: other desired targets and the stdin controller continue running. Including that target in a later complete set starts a fresh generation.

EXIT STATUS

0 on success (including a finite run's ordinary completion, and an interrupted run of either mode). Nonzero if the managed driver failed, if an uninterrupted continuous run ended by peer closure, or if the run was not interrupted and no target completed successfully. An interrupted run never fails on the last of these conditions — for example, interrupting a continuous run whose only target never opened still exits 0.

SERVER FILL REQUEST

--sfill DESCRIPTOR requests the server's echo-reply payload fill: none (zero-filled), rand (random bytes), or pattern:HEX (a repeating hexadecimal pattern). See irtt-server(1) SERVER FILL for what the server does with an unset or unparseable request.

EXAMPLES

irtt-client netperf-eu.bufferbloat.net:2112
irtt-client netperf-eu.bufferbloat.net:2112 --duration 30s --interval 100ms
irtt-client netperf-eu.bufferbloat.net:2112 --duration 0
irtt-client host-a:2112 host-b:2112 --pacing burst

SEE ALSO

irtt-tui(1), irtt-server(1), irtt-rs(1), and irtt-client --help for the full option list, including HMAC, clock/timestamp negotiation, DSCP, and TTL flags.