18975 FIX ibm_mq_queues: Support QStatus from CSQM format

Users monitoring IBM MQ on Mainframe (z/OS) systems via Piggyback
encountered crashes and missing data following the changes in Werk

The following problems have been resolved:
Crash: "ValueError: Cannot render negative timespan": `ibm_mq_queues`
check crashed with a `ValueError` when the IBM MQ system clock was ahead
or due to parsing precision differences. Use absolute time difference to
prevent negative timespans and crashes.

Data Merging: Users reported `duplicate section definition` errors or
missing service details when monitoring Mainframe queues. The agent
receives split information for `QUEUE` and `QSTATUS`.

The plugin now parses and merges `QUEUE` and `QSTATUS` outputs
correctly. This merge combines the mq data into a single service. This
resolves the crash and populates all metrics correctly.

Parsing of Empty Values: The parser now handles empty values in the
agent output correctly. Previously, strict parsing logic led to
incomplete data parsing or missing services.

Before this fix, you might have seen crashes in the `ibm_mq_queues`
check or missing queue metrics. After updating, monitoring will be
stable and complete without needing any changes to your configuration.

JIRA-Ref: SUP-23371 | SUP-26328 | CMK-32551
Change-Id: If818fecfd866aa24c80eb45eb32284ab0f12c958

Author: smeagol91

.werks/18975.md

@@ -0,0 +1,26 @@
+[//]: # (werk v2)
+# ibm_mq_queues: Support additional QStatus informations from CSQM format
+
+key        | value
+---------- | ---
+date       | 2025-12-15T10:00:36+00:00
+version    | 2.3.0p46
+class      | fix
+edition    | cre
+component  | checks
+level      | 1
+compatible | yes
+
+Users monitoring IBM MQ on Mainframe (z/OS) systems via Piggyback encountered crashes and missing data following the changes in Werk #18303.
+
+The following problems have been resolved:
+
+**Crash: "ValueError: Cannot render negative timespan"**: `ibm_mq_queues` check crashed with a `ValueError` when the IBM MQ system clock was ahead or due to parsing precision differences. Use absolute time difference to prevent negative timespans and crashes.
+
+**Data Merging**: Users reported `duplicate section definition` errors or missing service details when monitoring Mainframe queues. The agent receives split information for `QUEUE` and `QSTATUS`.
+
+The plugin now parses and merges `QUEUE` and `QSTATUS` outputs correctly. This merge combines the mq data into a single service. This resolves the crash and populates all metrics correctly.
+
+**Parsing of Empty Values**: The parser now handles empty values in the agent output correctly. Previously, strict parsing logic led to incomplete data parsing or missing services.
+
+Before this fix, you might have seen crashes in the `ibm_mq_queues` check or missing queue metrics. After updating, monitoring will be stable and complete without needing any changes to your configuration.

cmk/base/legacy_checks/ibm_mq_queues.py

@@ -161,7 +161,7 @@ def ibm_mq_last_age(mq_date, mq_time, agent_timestamp, label, key, params):
         return (0, label + ": n/a", [])
     mq_datetime = "{} {}".format(mq_date, mq_time.replace(".", ":"))
     input_time = dateutil.parser.parse(mq_datetime, default=agent_timestamp)
-    age = (agent_timestamp - input_time).total_seconds()
+    age = abs((agent_timestamp - input_time).total_seconds())
     return check_levels(
         age, None, params.get(key), human_readable_func=render.timespan, infoname=label
     )

cmk/base/plugins/agent_based/ibm_mq_queues.py

@@ -15,7 +15,19 @@
 
 
 def parse_ibm_mq_queues(string_table: StringTable) -> Section:
-    return parse_ibm_mq(string_table, "QUEUE")
+    queue_status = parse_ibm_mq(string_table, "QSTATUS")
+    queue = parse_ibm_mq(string_table, "QUEUE")
+    return _deep_merge_sections(queue, queue_status)
+
+
+def _deep_merge_sections(priority_section: Section, additional_section: Section) -> Section:
+    result = dict(additional_section)
+    for key, value in priority_section.items():
+        if key in result and isinstance(result[key], dict) and isinstance(value, dict):
+            result[key] = _deep_merge_sections(value, result[key])
+        else:
+            result[key] = value
+    return result
 
 
 register.agent_section(

cmk/plugins/lib/ibm_mq.py

@@ -14,8 +14,8 @@
 RE_INTRO = re.compile(r"^QMNAME\((.*)\)[\s]*STATUS\((.*?)\)[\s]*NOW\((.*)\)")
 RE_GROUP = re.compile(r"^(AMQ\d+\w?:|CSQM\d+\w? !)")
 RE_KEY = re.compile(r"([\s]*|CSQM\d+\w? ![A-Z0-9.]+[\s]+)([A-Z0-9]+\()")
-RE_SECOND_COLUMN = re.compile(r"[A-Z]+\([^&]+\)$")
-RE_KEY_VALUE = re.compile(r"([A-Z0-9]+)\(([^&]+)\)")
+RE_SECOND_COLUMN = re.compile(r"[A-Z]+\([^&]*\)$")
+RE_KEY_VALUE = re.compile(r"([A-Z0-9]+)\(([^&]*)\)")
 
 
 def parse_ibm_mq(string_table: StringTable, group_by_object: str) -> Section: