Linux default checks and policies

  • リリースバージョン: Australia
  • 更新日 2026年03月12日
  • 所要時間：32分

    • Agent Client Collector provides the following default checks and policies for Linux Metrics monitoring.

    Linux monitoring metrics checks

    表 : 1. Linux metrics policy
    Check Metric Name Resource Description Units Featured Metric Anomaly Detection
    os.linux.metrics-process-usage proc.acc.running process-name Number of processes running with this name(acc) count
    proc.acc.cpuPercent process-name Percentage of cpu taken by the process. percent

    proc.acc.​memPercent

    		 process-name Percentage of memory taken by the process. percent
    os.linux.metrics-reboot-count-today reboot.count.today empty Number of reboot done on today count
    os.linux.metrics-system-cpu cpu.total.user total Normal processes executing in user mode; cpu.total.user is the total of the cpuN.user metrics. count
    cpu.total.nice total Niced processes executing in user mode; cpu.total.nice is the total of the cpuN.nice metrics. count
    cpu.total.system total Time the CPU spent running the kernel; cpu.total.system is the total of the cpuN.system metrics. sec
    cpu.total.idle total Total time the CPU spent in an idle state.; cpu.total.idle is the total of the cpuN.idle metrics. sec
    cpu.total.iowait total Total time the CPU spent waiting for IO operations to complete.; cpu.total.iowait is the total of the cpuN.iowait metrics. sec
    cpu.total.irq total Total time that the processor is spending on handling Interrupts.; cpu.total.irq is the total of the cpuN.irq metrics. sec
    cpu.total.softirq total Time spent on servicing soft interrupt requests; cpu.total.softirq is the total of the cpuN.softirq metrics. sec
    cpu.total.steal total Total time the virtual CPU spent waiting for the hypervisor to service another virtual CPU. Only applies to virtual machines. sec
    cpu.total.guest total Total time the CPU spent running the virtual processor. Only applies to hypervisors. sec
    cpu.total.guest_nice total Total time the CPU spent running as nice guest OS. cpu.total.guset_nice si the total of the cpuN.guest_nice metrics sec
    cpu.<cpu-core>.user cpu-core Time spent with normal processing in user mode. sec
    cpu.<cpu-core>.nice cpu-core Time spent with niced processes in user mode. sec
    cpu.<cpu-core>.system cpu-core Time spent running in kernel mode. sec
    cpu.<cpu-core>.idle cpu-core Time spent in vacations twiddling thumbs. sec
    cpu.<cpu-core>.iowait cpu-core Time spent waiting for I/O to completed. This is considered idle time too. sec
    cpu.<cpu-core>.irq cpu-core Time spent serving hardware interrupts. sec
    cpu.<cpu-core>.softirq cpu-core Time spent serving software interrupts. sec
    cpu.<cpu-core>.steal cpu-core Time stolen by other operating systems running in a virtual environment. sec
    cpu.<cpu-core>.guest cpu-core Time spent for running a virtual CPU or guest OS under the control of the kernel. sec
    cpu.<cpu-core>.guest_nice cpu-core Total time the CPU spent running as nice guest OS. sec
    cpu.intr empty Interrupts serviced since boot time. sec
    cpu.ctxt empty Total number of context switches across all CPUs. count
    cpu.btime empty The time the system booted sec
    cpu.processes empty The number of processes and threads created, which includes (but is not limited to) those created by calls to the fork() and clone() system calls. count
    cpu.procs_running empty The total number of processes running on all CPUs. count
    cpu.procs_blocked empty The number of processes currently blocked, waiting for I/O to complete. count
    cpu.cpu_count empty Number of cpu on the system count
    cpu.<cpu-core>.cores cpu-core The number of CPU cores core count
    os.linux.metrics-system-cpu-load load_avg.one empty The average system load over one minute. thread count yes yes
    load_avg.five empty The average system load over five minutes. thread count yes yes
    load_avg.fifteen empty The average system load over fifteen minutes. thread count yes yes
    load_avg.norm.one empty The average system load over one minute normalized by the number of CPUs. thread count
    load_avg.norm.five empty The average system load over five minutes normalized by the number of CPUs. thread count
    load_avg.norm.fifteen empty The average system load over fifteen minutes normalized by the number of CPUs. thread count
    os.linux.metrics-system-cpu-percentage cpu.avgutilization_​percentage empty Percent of cpu was used on average percent
    cpu.user_​percentage empty Percent of time total cpu was used by normal processes in user mode percent yes yes
    cpu.nice_​percentage empty Percent of time all cpus used by niced processes in user mode percent yes yes

    cpu.system_​percentage

    		 empty The percent of time the CPU spent running the kernel. percent yes yes
    cpu.idle_percentage empty Percent of time all cpus were idle percent yes yes

    cpu.iowait_​percentage

    		 empty Percent of time all cpus waiting for I/O to complete percent yes yes
    cpu.irq_percentage empty Percent of time all cpus servicing interrupts percent yes yes

    cpu.softirq_​percentage

    		 empty Percent of time all cpus servicing software interrupts percent yes yes
    cpu.steal_​percentage empty Percent of time all cpus serviced virtual hosts operating systems percent yes yes
    cpu.guest_​percentage empty Percent of time all cpus serviced guest operating system percent yes yes
    os.linux.metrics-system-disk disk.<disk-name>.reads disk-name Total number of reads completed successfully. count yes yes
    disk.<disk-name>.readsMerged disk-name Total number of reads merged count
    disk.<disk-name>.sectorsRead disk-name Total number of sectors read successfully. count
    disk.<disk-name>.readTime disk-name Total number of milliseconds spent by all reads. millisec
    disk.<disk-name>.writes disk-name Total number of writes completed successfully. count yes yes
    disk.<disk-name>.writesMerged disk-name Total number of writes merged count
    disk.<disk-name>.sectorsWritten disk-name Total number of sectors written successfully. count
    disk.<disk-name>.writeTime disk-name Total number of milliseconds spent by all writes. misllisec
    disk.<disk-name>.ioInProgress disk-name Total number of I/Os currently in progress count
    disk.<disk-name>.ioTime Total time spent doing I/Os millisec yes yes
    disk.<disk-name>.ioTimeWeighted disk-name Total time spent doing I/Os. This can provide an easy measure of both I/O completion time and the backlog that may be accumulating. millisec
    os.linux.metrics-system-disk-capacity disk.<file-system-name>.total file-system-name The total size of the file system. byte
    disk.<file-system-name>.used file-system-name The total amount of space allocated to existing files in the file system. byte
    disk.<file-system-name>.avail file-system-name The total amount of space available within the file system. byte
    disk.<file-system-name>.used_percentage file-system-name The percentage of the available space that currently allocated to all files on the file system. percent
    disk.<file-system-name>.itotal file-system-name The total number of inodes on the file system. count
    disk.<file-system-name>.iused file-system-name The number of used inodes. count
    disk.<file-system-name>.iavail file-system-name The number of free (unused) inodes. count
    disk.<file-system-name>.iused_percentage file-system-name The percentage of used inodes. percent
    os.linux.metrics-system-disk-usage disk_usage.<disk>.total disk-name Total amount of space available on this disk bytes
    disk_usage.<disk>.used disk-name Total amount of space used in this disk bytes
    disk_usage.<disk>.avail disk-name Total amount of space available on this disk bytes
    disk_usage.<disk>.used_​percentage disk-name The percentage of space used on this disk percent yes yes
    os.linux.metrics-system-memoryos.​linux.metrics-system-​memory-percent memory.total empty Total usable RAM. KB
    memory.free empty Total free RAM. KB
    memory.available empty An estimate of how much memory is available for starting new applications, without swapping. KB
    memory.buffers empty Temporary storage used for raw disk blocks. KB
    memory.cached empty In-memory cache for files read from disk (the page cache). Does not include mem_swapcached. KB
    memory.swapTotal empty Total amount of swap space available. KB yes yes
    memory.swapFree empty Amount of swap space that is currently unused. yes yes
    memory.dirty empty Memory which is waiting to get written back to the disk. KB
    memory.swapUsed empty The amount of swap space in use. KB yes yes
    memory.used empty The amount of RAM in use. KB
    memory.​usedWOBuffersCaches empty The amount of memory in use. KB
    memory.​freeWOBuffersCaches empty Value of MemAvailable from /proc/meminfo if present, but falls back to adding free + buffered + cached memory if not. KB
    memory.​swapUsedPercentage empty Percent of swap space used. percent
    memory_percent.​free empty Percent of free RAM percent yes yes
    memory_percent.​available empty Percent of Mem available percent yes yes
    memory_percent.​buffers empty Precent of Memory used for raw disk blocks percent yes yes
    memory_percent.​cached empty Percent of memory used for in-memory cache for files read from disk percent yes yes
    memory_percent.​dirty empty Percent of memory waiting to get written back to the disk. percent yes yes
    memory_percent.​swapUsed empty Percent of swap space used. percent yes yes
    memory_percent.​usedWOBuffersCaches empty Percent of memory is being used percent yes yes
    memory_percent.​freeWOBuffersCaches empty Percent of Memory available percent yes yes
    os.linux.metrics-​system-uptime system.uptime(sec) empty The amount of time the system has been working and available. sec
    os.linux.metrics-​memory-vmstat vmstat.nr_free_pages empty Pages that are currently unused by the system. pages
    vmstat.nr_alloc_​batch empty pages allocated to other domains due to insufficient memory in each domain of each NUMA pages
    vmstat.nr_​inactive_anon empty memory pages in each domain of each NUMA node that have not been accessed for a long time pages
    vmstat.nr_active_​anon empty Anonymous virtual memory pages that have been recently used KB
    vmstat.nr_inactive_​file empty The memory page corresponding to the file that has not been accessed for a long time in each domain of each NUMA. KB
    vmstat.nr_active_​file empty The memory page corresponding to the file that has been accesseed recently . pages
    vmstat.nr_​unevictable empty The number of pages is in the unevictable (non-)LRU list count
    vmstat.nr_​mlock empty Pages mapped into a VM_LOCKED VMA - are a class of unevictable pages. pages
    vmstat.nr_anon_​pages empty Memory mapped pages that is not part of a file. pages
    vmstat.nr_​mapped empty The number of memory mapped pages. count
    vmstat.nr_file_​pages empty
    vmstat.nr_​dirty empty Pages waiting to be written to disk pages
    vmstat.nr_​writeback empty Pages currently being written to disk pages
    vmstat.nr_slab_​reclaimable empty Pages from the kernel slab memory usage that can be reclaimed pages
    vmstat.nr_slab_​unreclaimable empty Pages from the kernel slab memory usage that cannot be reclaimed pages
    vmstat.nr_page_table_​pages empty Pages allocated to page tables pages
    vmstat.nr_kernel_​stack empty Amount of memory allocated to kernel stacks. KB
    vmstat.nr_unstable empty The number of unstable pages in each domain of each NUMA node count
    vmstat.nr_bounce empty
    vmstat.nr_vmscan_​write empty The number of dirty pages written back during a scan of LRU(s) count
    vmstat.nr_vmscan_​immediate_reclaim empty
    vmstat.nr_writeback_​temp empty
    vmstat.nr_isolated_​anon empty The number of anonymous memory pages isolated in each domain of each NUMA node count
    vmstat.nr_​isolated_file empty The number of pages of file storage pages isolated in each domain of each NUMA node count
    vmstat.nr_shmem empty The number of shared memory pages count
    vmstat.nr_dirtied empty The number of dirty pages in each domain of each NUMA node count
    vmstat.nr_written empty
    vmstat.numa_hit empty The number of pages that were successfully allocated to this node. count
    vmstat.numa_miss empty The number of pages that were allocated on this node because of low memory on the intended node. count
    vmstat.numa_​foreign empty The number of pages initially intended for this node that were allocated to another node instead. count
    vmstat.numa_​interleave empty The number of interleave policy pages successfully allocated to this node. count
    vmstat.numa_local empty The number of pages successfully allocated on this node, by a process on this node count
    vmstat.numa_other empty The number of pages allocated on this node, by a process on another node. count
    vmstat.workingset_​refault empty
    vmstat.workingset_​activate empty
    vmstat.workingset_​nodereclaim empty
    vmstat.nr_anon_transparent_​hugepages empty
    vmstat.nr_free_cma empty Free continuous memory allocator pages in each domain of each NUMA
    vmstat.nr_dirty_​threshold empty
    vmstat.nr_dirty_​background_threshold empty
    vmstat.pgpgin empty The number of pages brought in from disk count
    vmstat.pgpgout empty The number of pages written out to disk count
    vmstat.pswpin empty The number of pages brought in from swap space count
    vmstat.pswpout empty The number of pages swapped out into swap space count
    vmstat.pgalloc_dma empty
    vmstat.pgalloc_​dma32 empty
    vmstat.pgalloc_​normal empty
    vmstat.pgalloc_​movable empty
    vmstat.pgfree empty The number of pages are free since last boot count
    vmstat.pgactivat empty Number of page activations since last boot count
    vmstat.pgdeactivate empty Number of page deactivations since last boot count
    vmstat.pgfault empty Minor faults since last boot pages
    vmstat.pgmajfault empty Major faults since last boot pages
    vmstat.pglazyfreed empty
    vmstat.pgrefill_dma empty
    vmstat.pgrefill_dma32 empty
    vmstat.pgrefill_normal empty Number of page refills since last boot count
    vmstat.pgrefill_​movable empty
    vmstat.pgsteal_​kswapd_dma empty
    vmstat.pgsteal_​kswapd_dma32 empty
    vmstat.pgsteal_​kswapd_normal empty
    vmstat.pgsteal_​kswapd_movable empty
    vmstat.pgsteal_​direct_dma empty
    vmstat.pgsteal_​direct_dma32 empty
    vmstat.pgsteal_​direct_normal empty
    vmstat.pgsteal_​direct_movable empty
    vmstat.pgscan_​kswapd_dma empty
    vmstat.pgscan_​kswapd_dma32 empty
    vmstat.pgscan_​kswapd_normal empty Number of pages scanned by kswapd since boot count
    vmstat.pgscan_​kswapd_movable empty
    vmstat.pgscan_​direct_dma empty
    vmstat.pgscan_​direct_dma32 empty
    vmstat.pgscan_​direct_normal empty Number of pages reclaimed since boot count
    vmstat.pgscan_​direct_movable empty
    vmstat.pgscan_​direct_throttle empty
    vmstat.zone_​reclaim_failed empty
    vmstat.pginodesteal empty
    vmstat.slabs_scanned empty
    vmstat.kswapd_​inodesteal empty
    vmstat.kswapd_low_​wmark_hit_quickly empty
    vmstat.kswapd_high_​wmark_hit_quickly empty
    vmstat.pageoutrun empty Number of times kswapd called page reclaim count
    vmstat.allocstall empty Number of times page reclaim was called directly (low memory) count
    vmstat.pgrotated empty
    vmstat.drop_​pagecache empty
    vmstat.drop_​slab empty
    vmstat.numa_pte_​updates empty
    vmstat.numa_huge_​pte_updates empty
    vmstat.numa_hint_​faults empty
    vmstat.numa_hint_​faults_local empty
    vmstat.numa_pages_​migrated empty
    vmstat.pgmigrate_​success empty
    vmstat.pgmigrate_fail empty
    vmstat.compact_​migrate_scanned empty
    vmstat.compact_​free_scanned empty
    vmstat.compact_​isolated empty
    vmstat.compact_​stall empty The number of times a process stalls to run memory compaction so that a huge page is free for use. count
    vmstat.compact_fail empty The number of times the system tries to compact memory but failed. count
    vmstat.compact_​success empty The number of times the system compacted memory and freed a huge page for use. count
    vmstat.htlb_buddy_​alloc_success empty
    vmstat.htlb_buddy_​alloc_fail empty
    vmstat.unevictable_​pgs_culled empty
    vmstat.unevictable_​pgs_scanned empty
    vmstat.unevictable_​pgs_rescued empty
    vmstat.unevictable_​pgs_mlocked empty
    vmstat.unevictable_​pgs_munlocked empty
    vmstat.unevictable_​pgs_cleared empty
    vmstat.unevictable_​pgs_stranded empty
    vmstat.thp_fault_alloc empty The number of huge pages is successfully allocated to handle a page fault. count
    vmstat.thp_fault_​fallback empty The number of page fault fails to allocate a huge page and instead falls back to using small pages. count
    vmstat.thp_collapse_alloc empty The number of collapse of a range of pages into one huge page and then successfull allocation of a new huge page to store the data. count
    vmstat.thp_collapse_​alloc_failed empty The number of collapse of a range of pages into one huge page but failed the allocation. count
    vmstat.thp_split empty The number of split of a huge page into base pages count
    vmstat.thp_zero_​page_alloc empty The number of successful allocation of huge zero page count
    vmstat.thp_zero_​page_alloc_failed empty The number of times the kernel failed to allocate huge zero page and falls back to using small pages. count
    vmstat.balloon_inflate empty
    vmstat.balloon_​deflate empty
    vmstat.balloon_​migrate empty
    os.linux.metrics-process-status proc.<process>.VmSize process-name The total amount of virtual memory used by the process KB
    proc.<process>.VmRSS process-name The non-swapped physical memory a process has used KB
    proc.<process>.VmSwap process-name The total amount of swap space used. KB

    Linux network monitoring checks

    注:
    When upgrading from an earlier version, manually add the checks in this table to the Linux metrics policy.
    表 : 2. Linux metrics policy
    Type Check Description Usage and usage example Metrics collected Featured Metric
    Metric os.linux.metrics-network-interface Retrieves all network interface related metrics for Linux servers.
    Usage:
    • -x, excludeinterface: List of interfaces to exclude (comma separated)
    • -i, includeinterface: List of interfaces to include (comma separated)
    • -I, includeinterfaceregex: Regex matching interfaces to include
    • -X, excludeinterfaceregex: Regex matching interfaces to exclude

    Usage example: command: metrics-network-interface.rb
    • rxBytes (featured metric)
    • rxPackets (featured metric)
    • rxErrors
    • rxDrops (featured metric)
    • rxFifo
    • rxFrame
    • rxCompressed
    • rxMulticast
    • rxBytes (featured metric)
    • rxPackets (featured metric)
    • rxErrors
    • rxDrops (featured metric)
    • rxFifo
    • rxColls
    • rxCarrier
    • rxCompressed
    		 yes
    Metric os.linux.metrics-netstat-tcp Retrieves metrics on TCP socket states from netstat. Useful on high-traffic web or proxy servers with large numbers of short-lived TCP connections coming and going.
    Usage:
    • -p, port: The port from which you want to receive metrics. Value range = 1-65535.
    • -t, type: The port type from which to receive metrics. Values=local or remote. Default = local
    • -d, disabletcp6: Disables the tcp6 check. Enter a value to set disabletcp6 = true.

    Usage example: metrics-netstat-tcp.rb
    • tcp.UNKNOWN
    • tcp.ESTABLISHED
    • tcp.SYN_SENT
    • tcp.SYN_RECV
    • tcp.FIN_WAIT1
    • tcp.FIN_WAIT2
    • tcp.TIME_WAIT
    • tcp.CLOSE
    • tcp.CLOSE_WAIT
    • tcp.LAST_ACK
    • tcp.LISTEN
    • tcp.CLOSING
    		 no