Uploaded image for project: 'RHEL'
  1. RHEL
  2. RHEL-106847

pcp2openmetrics reports only the first metric when uses pmlogger archives

Linking RHIVOS CVEs to...Migration: Automation ...Sync from "Extern...XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Unresolved
    • Icon: Normal Normal
    • None
    • rhel-9.7
    • pcp
    • No
    • Moderate
    • rhel-pt-pcp
    • 2
    • QE ack
    • False
    • False
    • Hide

      None

      Show
      None
    • No
    • Red Hat Enterprise Linux
    • None
    • None
    • Automated
    • Unspecified Release Note Type - Unknown
    • Unspecified
    • Unspecified
    • Unspecified
    • All
    • None

      The pcp2openmetrics tool returns only one (the first in the archive) if multiple metrics are specified on command line.

      Compose: RHEL-9.7.0-updates-20250729.2
      PCP: pcp-6.3.7-5.el9

      Requesting i.e. hinv.ncpu metric from an archive returns:

      $ pcp2openmetrics -a /var/log/pcp/pmlogger/127.0.0.2/20250731.08.34 -s 1 'hinv.ncpu'
# PCP5 hinv_ncpu 60.0.32 u32 PM_INDOM_NULL discrete 
# TYPE hinv_ncpu gauge
# HELP hinv_ncpu number of CPUs in the system
hinv_ncpu{domainname="localdomain",hostname="s390x-kvm-073",machineid="6db8b81d81754b269c1c1d4c12348d02",agent="linux"} 4 1753979678.632047
# EOF

      Requesting a different metric i.e. kernel.all.load from the same archive returns:

      $ pcp2openmetrics -a /var/log/pcp/pmlogger/127.0.0.2/20250731.08.34 -s 1 'kernel.all.load'
# PCP5 kernel_all_load 60.2.0 float 60.2 instant 
# TYPE kernel_all_load gauge
# HELP kernel_all_load 1, 5 and 15 minute load average
kernel_all_load{domainname="localdomain",hostname="s390x-kvm-073",machineid="6db8b81d81754b269c1c1d4c12348d02",instname="1 minute",instid="1",agent="linux"} 0.520 1753979679.122822
kernel_all_load{domainname="localdomain",hostname="s390x-kvm-073",machineid="6db8b81d81754b269c1c1d4c12348d02",instname="5 minute",instid="5",agent="linux"} 0.250 1753979679.122822
kernel_all_load{domainname="localdomain",hostname="s390x-kvm-073",machineid="6db8b81d81754b269c1c1d4c12348d02",instname="15 minute",instid="15",agent="linux"} 0.140 1753979679.122822
# EOF

      However when both the metrics above are requested, only the one which is found first in the archive is returned:

      $ pcp2openmetrics -a /var/log/pcp/pmlogger/127.0.0.2/20250731.08.34 -s 1 'kernel.all.load' 'hinv.ncpu'
# PCP5 hinv_ncpu 60.0.32 u32 PM_INDOM_NULL discrete 
# TYPE hinv_ncpu gauge
# HELP hinv_ncpu number of CPUs in the system
hinv_ncpu{domainname="localdomain",hostname="s390x-kvm-073",machineid="6db8b81d81754b269c1c1d4c12348d02",agent="linux"} 4 1753979678.632047
# EOF

      In comparission, when the pcp2openmetrics tool is used against pmcd (not an archive), it returns both metrics as expected:

      $ pcp2openmetrics -s 1 'kernel.all.load' 'hinv.ncpu'
# PCP5 kernel_all_load 60.2.0 float 60.2 instant 
# TYPE kernel_all_load gauge
# HELP kernel_all_load 1, 5 and 15 minute load average
kernel_all_load{domainname="localdomain",groupid="0",hostname="s390x-kvm-073",machineid="6db8b81d81754b269c1c1d4c12348d02",userid="0",instname="1 minute",instid="1",agent="linux"} 0.000
kernel_all_load{domainname="localdomain",groupid="0",hostname="s390x-kvm-073",machineid="6db8b81d81754b269c1c1d4c12348d02",userid="0",instname="5 minute",instid="5",agent="linux"} 0.000
kernel_all_load{domainname="localdomain",groupid="0",hostname="s390x-kvm-073",machineid="6db8b81d81754b269c1c1d4c12348d02",userid="0",instname="15 minute",instid="15",agent="linux"} 0.000
hinv_ncpu{domainname="localdomain",groupid="0",hostname="s390x-kvm-073",machineid="6db8b81d81754b269c1c1d4c12348d02",userid="0",agent="linux"} 4
# EOF

              pcp-maint pcp-maint
              jkurik@redhat.com Jan Kurik
              pcp-maint pcp-maint
              Jan Kurik Jan Kurik
              Votes:
              0 Vote for this issue
              Watchers:
              5 Start watching this issue

                Created:
                Updated: