feat(disk-usage, dhcp-scope-usage): add --brief switch to hide rows within thresholds

--brief is the row-level counterpart of the column-level --lengthy
switch: --lengthy adds columns, --brief filters rows. They are
orthogonal and both can be set at the same time. Perfdata and
alerting are strictly unaffected: every checked item still emits
perfdata so Grafana trending stays complete, and every item still
drives the overall check state.

When --brief is set and no item is in WARN/CRIT state, the plugin
collapses to a single-line "Everything is ok. (thresholds)" summary
without an empty table.

Admins on hosts with hundreds of disk mounts or thousands of DHCP
scopes can now use --brief to get event-console-sized output while
keeping the full perfdata for Grafana. The convention is documented
in CONTRIBUTING under "Verbosity parameter convention"; #1081 tracks
the sweep to extend --brief to other plugins with unbounded table
output.

Also picks up regenerated Icinga Director basket files for
dhcp-scope-usage, disk-usage, docker-info, haproxy-status and
podman-info (the latter three from recent --ignore work).

Closes #782
Closes #788

Author: markuslf

CHANGELOG.md

@@ -33,6 +33,8 @@ Monitoring Plugins:
 
 * by-ssh: add alerting on single numeric values
 * by-winrm: executes commands on remote Windows hosts by WinRM, supporting JEA (including the JEA endpoint via `--winrm-configuration-name`)
+* dhcp-scope-usage: add `--brief` parameter to hide scopes within the thresholds, useful on DHCP servers with thousands of scopes where the default output becomes unreadable. Inverse of `--lengthy`: `--brief` filters rows, `--lengthy` adds columns, both are orthogonal and can be combined. Perfdata and alerting are unaffected, so Grafana trending stays complete ([#788](https://github.com/Linuxfabrik/monitoring-plugins/issues/788))
+* disk-usage: add `--brief` parameter to hide filesystems within the thresholds, useful on hosts with hundreds of mounts where the default output becomes unreadable. Inverse of `--lengthy`: `--brief` filters rows, `--lengthy` adds columns, both are orthogonal and can be combined. Perfdata and alerting are unaffected, so Grafana trending stays complete ([#782](https://github.com/Linuxfabrik/monitoring-plugins/issues/782))
 * infomaniak-swiss-backup-devices: add `--ignore-customer`, `--ignore-name`, `--ignore-tag`, `--ignore-user` parameters to skip devices by regex
 * infomaniak-swiss-backup-products: add `--ignore-customer`, `--ignore-tag` parameters to skip products by regex
 * docker-info, podman-info: add `--ignore` parameter to filter stderr warnings and errors by regex, e.g. to suppress the `WARNING: No swap limit support` message from `docker info` on kernels without swap accounting, or the `WARNING: bridge-nf-call-iptables is disabled` message on Debian hosts that do not load the `br_netfilter` module. Accepts a Python regex, is case-sensitive by default (use `(?i)` for case-insensitive matching) and can be specified multiple times. The same treatment was applied to both plugins because they share the same stderr-parsing code path ([#834](https://github.com/Linuxfabrik/monitoring-plugins/issues/834))

CONTRIBUTING.md

@@ -138,7 +138,7 @@ Checklist:
 
 ### Rules of Thumb
 
-* Be brief by default. Report what needs to be reported to fix a problem. If there is more information that might help the admin, support a `--lengthy` parameter.
+* Be brief by default. Report what needs to be reported to fix a problem. If there is more information that might help the admin, support a `--lengthy` parameter. If the default output still grows unbounded on large systems (thousands of disk mounts, DHCP scopes, backends, services), also support a `--brief` parameter that hides rows within the thresholds. See "Verbosity parameter convention" below.
 * The plugin should be "self configuring" and/or using best practise defaults, so that it runs without parameters wherever possible.
 * Develop with a minimal Linux in mind.
 * Develop with Icinga2 in mind.
@@ -232,6 +232,7 @@ For all other options, use long parameters only. Separate words using a `-`. We
 --always-ok
 --argument
 --authtype
+--brief
 --cache-expire
 --command
 --community
@@ -262,8 +263,6 @@ For all other options, use long parameters only. Separate words using a `-`. We
 --icinga-username
 --idsite
 --ignore
---ignore-pattern
---ignore-regex
 --input
 --insecure
 --instance
@@ -600,6 +599,27 @@ If a plugin supports `-v`/`--verbose`, it should implement up to three verbosity
 Note: Most of our plugins use `--lengthy` instead of `-v` for extended output. The verbosity levels above apply if the plugin explicitly supports `--verbose`.
 
 
+### Verbosity parameter convention: `--lengthy` and `--brief`
+
+`--lengthy` and `--brief` are the two verbosity knobs admins use to tune what a plugin prints. They are **orthogonal** (not mutually exclusive) and control different axes of the output:
+
+| Parameter | Axis | Effect |
+|-----------|------|--------|
+| default | rows × columns | Show all checked items with the core columns. |
+| `--lengthy` | columns | **Add** extra columns to every row (e.g. full details, debug info). |
+| `--brief` | rows | **Hide** rows that are within the thresholds. Show only items in WARN/CRIT state. |
+| `--lengthy --brief` | rows × columns | Hide OK rows, show extra columns on the rows that remain. |
+
+Rules:
+
+* **Perfdata is always complete.** `--brief` and `--lengthy` only reshape the human-readable message. Every checked item still emits perfdata so Grafana can trend everything.
+* **Alerting is unaffected.** All items (including the ones `--brief` hides) still drive the overall check state. `--brief` is a display filter, not a threshold.
+* **When `--brief` hides everything**, the plugin prints only the summary header ("Everything is ok. (thresholds)"), not an empty table. Admins on a quiet system see one line.
+* **`--lengthy` and `--brief` are always combinable.** Do not mark them mutually exclusive in `argparse`.
+* **When to support `--brief`**: add it whenever the default output can grow unbounded on large systems (hundreds of disk mounts, thousands of DHCP scopes, hundreds of HAProxy backends, etc.). Reference implementations: `check-plugins/disk-usage` and `check-plugins/dhcp-scope-usage`.
+* **Help text** for `--brief` should describe the filter semantic and explicitly state that perfdata and alerting are unaffected, so the admin understands that `--brief` is safe to use on production without losing trending data.
+
+
 ### Plugin Performance Data, Perfdata
 
 "UOM" means "Unit of Measurement".

check-plugins/dhcp-scope-usage/README.md

@@ -34,22 +34,30 @@ Monitors IPv4 DHCP scope usage on a Windows DHCP server. Connects via WinRM and
 ## Help
 
 ```text
-usage: dhcp-scope-usage [-h] [-V] [--always-ok] [-c CRIT] [-H HOSTNAME]
-                        [--test TEST] [-w WARN] [--winrm-domain WINRM_DOMAIN]
+usage: dhcp-scope-usage [-h] [-V] [--always-ok] [--brief] [-c CRIT]
+                        [-H HOSTNAME] [--test TEST] [-w WARN]
+                        [--winrm-domain WINRM_DOMAIN]
                         --winrm-hostname WINRM_HOSTNAME
                         --winrm-password WINRM_PASSWORD
                         [--winrm-transport {basic,ntlm,kerberos,credssp,plaintext}]
                         [--winrm-username WINRM_USERNAME]
 
 Monitors IPv4 DHCP scope usage on a Windows DHCP server. Connects via WinRM
-and queries scope statistics using PowerShell. Alerts when the address pool
-usage of any scope exceeds the configured thresholds (default: WARN at 80%,
-CRIT at 90%).
+and queries scope statistics using PowerShell. On servers with thousands of
+scopes, --brief hides rows within the thresholds so the output only lists
+scopes in WARN/CRIT state. Alerts when the address pool usage of any scope
+exceeds the configured thresholds (default: WARN at 80%, CRIT at 90%).
 
 options:
   -h, --help            show this help message and exit
   -V, --version         show program's version number and exit
   --always-ok           Always returns OK.
+  --brief               Hide scopes within the thresholds and show only those
+                        in WARN/CRIT state. Inverse of `--lengthy` (which adds
+                        columns); `--brief` filters rows. The two are
+                        orthogonal and can be combined. Perfdata and alerting
+                        are unaffected: all scopes still emit perfdata and
+                        still drive the overall check state. Default: False
   -c, --critical CRIT   CRIT threshold in percent. Default: >= 90
   -H, --hostname HOSTNAME
                         DNS name, IPv4, or IPv6 address of the DHCP server.

check-plugins/dhcp-scope-usage/dhcp-scope-usage

@@ -22,11 +22,13 @@ import lib.winrm
 from lib.globals import STATE_CRIT, STATE_OK, STATE_UNKNOWN, STATE_WARN
 
 __author__ = 'Linuxfabrik GmbH, Zurich/Switzerland'
-__version__ = '2026040801'
+__version__ = '2026041405'
 
 DESCRIPTION = """Monitors IPv4 DHCP scope usage on a Windows DHCP server. Connects via WinRM and
-queries scope statistics using PowerShell. Alerts when the address pool usage of any
-scope exceeds the configured thresholds (default: WARN at 80%, CRIT at 90%)."""
+queries scope statistics using PowerShell. On servers with thousands of scopes, --brief
+hides rows within the thresholds so the output only lists scopes in WARN/CRIT state.
+Alerts when the address pool usage of any scope exceeds the configured thresholds
+(default: WARN at 80%, CRIT at 90%)."""
 
 DEFAULT_CRIT = 90
 DEFAULT_HOSTNAME = 'localhost'
@@ -55,6 +57,18 @@ def parse_args():
         default=False,
     )
 
+    parser.add_argument(
+        '--brief',
+        help='Hide scopes within the thresholds and show only those in WARN/CRIT state. '
+        'Inverse of `--lengthy` (which adds columns); `--brief` filters rows. The two '
+        'are orthogonal and can be combined. Perfdata and alerting are unaffected: all '
+        'scopes still emit perfdata and still drive the overall check state. '
+        'Default: %(default)s',
+        dest='BRIEF',
+        action='store_true',
+        default=False,
+    )
+
     parser.add_argument(
         '-c',
         '--critical',
@@ -158,15 +172,15 @@ def main():
         lib.base.oao(stderr, STATE_WARN)
 
     # init some vars
-    msg = ''
     state = STATE_OK
     perfdata = ''
+    rows = []
 
     # ScopeId          Free            InUse           PercentageInUse  Reserved        Pending         SuperscopeName
     # -------          ----            -----           ---------------  --------        -------         --------------
     # 192.0.2.120      3               0               0                0               0
 
-    # get state and build the message
+    # get state and collect per-scope rows
     data_section = False
     for line in stdout.splitlines():
         # ignore all header lines including "----"
@@ -187,8 +201,8 @@ def main():
         scope_state = lib.base.get_state(scope_used, args.WARN, args.CRIT)
         state = lib.base.get_worst(scope_state, state)
 
-        # build the message
-        msg += f'* {scope_id}: {scope_used}% used{lib.base.state2str(scope_state, prefix=" ")}\n'
+        # Perfdata is emitted for every scope regardless of --brief so
+        # Grafana sees the full picture and trend lines stay continuous.
         perfdata += lib.base.get_perfdata(
             f'scope_{scope_id}',
             scope_used,
@@ -198,14 +212,32 @@ def main():
             _min=0,
             _max=100,
         )
+        rows.append((scope_id, scope_used, scope_state))
+
+    # Filter rows for --brief display. --brief is the inverse of
+    # --lengthy: --lengthy adds columns, --brief filters rows. They
+    # are orthogonal and both can be set at the same time. Perfdata
+    # and alerting stay untouched above this point; --brief only
+    # reshapes the human-readable output.
+    display_rows = (
+        [row for row in rows if row[2] != STATE_OK] if args.BRIEF else rows
+    )
+
+    # build the message
+    headers = {
+        STATE_CRIT: 'There are one or more criticals.',
+        STATE_WARN: 'There are one or more warnings.',
+        STATE_OK: 'Everything is ok.',
+    }
+    header = headers.get(state, headers[STATE_OK])
+    body = ''.join(
+        f'* {scope_id}: {scope_used}% used'
+        f'{lib.base.state2str(scope_state, prefix=" ")}\n'
+        for scope_id, scope_used, scope_state in display_rows
+    )
+    msg = f'{header}\n\n{body}' if body else header
 
     # over and out
-    if state == STATE_CRIT:
-        msg = 'There are one or more criticals.\n\n' + msg
-    elif state == STATE_WARN:
-        msg = 'There are one or more warnings.\n\n' + msg
-    else:
-        msg = 'Everything is ok.\n\n' + msg
     lib.base.oao(msg, state, perfdata, always_ok=args.ALWAYS_OK)