# Amazon Linux 2 This page describes the Optimization Pack for the component type Amazon Linux 2. ## Metrics ### CPU | Metric | | Description | | ------------------ | ------- | ----------------------------------------------------------------------------------------------------------------- | | cpu\_load\_avg | tasks | The system load average (i.e., the number of active tasks in the system) | | cpu\_num | CPUs | The number of CPUs available in the system (physical and logical) | | cpu\_util | percent | The average CPU utilization % across all the CPUs (i.e., how much time on average the CPUs are busy doing work) | | cpu\_used | CPUs | The average number of CPUs used in the system (physical and logical) | | cpu\_util\_details | percent | The average CPU utilization % broken down by usage type and CPU number (e.g., cp1 user, cp2 system, cp3 soft-irq) | ### Memory | Metric | | Description | | ------------------ | -------- | -------------------------------------------------------------------------------------------------------------- | | mem\_fault | faults/s | The number of memory faults (minor+major) | | mem\_fault\_major | faults/s | The number of major memory faults (i.e., faults that cause disk access) per second | | mem\_fault\_minor | faults/s | The number of minor memory faults (i.e., faults that do not cause disk access) per second | | mem\_swapins | pages/s | The number of memory pages swapped in per second | | mem\_swapouts | pages/s | The number of memory pages swapped out per second | | mem\_total | bytes | The total amount of installed memory | | mem\_used | bytes | The total amount of memory used | | mem\_used\_nocache | bytes | The total amount of memory used without considering memory reserved for caching purposes | | mem\_util | percent | The memory utilization % (i.e, the % of memory used) | | mem\_util\_details | percent | The memory utilization % (i.e., the % of memory used) broken down by usage type (e.g., active memory) | | mem\_util\_nocache | percent | The memory utilization % (i.e., the % of memory used) without considering memory reserved for caching purposes | ### Disk & Filesystem | Metric | | Description | | ----------------------------- | ------- | -------------------------------------------------------------------------------------------------------------------------------------------- | | disk\_io\_inflight\_details | ops | The number of IO disk operations in progress (outstanding) broken down by disk (e.g., disk /dev/nvme01) | | disk\_iops | ops/s | The average number of IO disk operations per second across all disks | | disk\_iops\_details | ops/s | The number of IO disk-write operations per second broken down by disk (e.g., disk /dev/nvme01) | | disk\_iops\_reads | ops/s | The average number of IO disk-read operations per second across all disks | | disk\_iops\_writes | ops/s | The average number of IO disk-write operations per second across all disks | | disk\_read\_bytes | bytes/s | The number of bytes per second read across all disks | | disk\_read\_bytes\_details | bytes/s | The average response time of IO disk operations broken down by disk (e.g., disk C://) | | disk\_read\_write\_bytes | bytes/s | The number of bytes per second written across all disks | | disk\_response\_time\_details | seconds | The average response time of IO disk operations broken down by disk (e.g., disk C://) | | disk\_response\_time\_read | seconds | The average response time of read disk operations | | disk\_response\_time\_worst | seconds | The average response time of IO disk operations of the slowest disk | | disk\_response\_time\_write | seconds | The average response time of write on disk operations | | disk\_swap\_used | bytes | The total amount of space used by swap disks | | disk\_swap\_util | percent | The average space utilization % of swap disks | | disk\_util\_details | percent | The utilization % of disk, i.e how much time a disk is busy doing work broken down by disk (e.g., disk D://) | | disk\_write\_bytes | bytes/s | The number of bytes per second written across all disks | | disk\_write\_bytes\_details | bytes/s | The number of bytes per second written from the disks broken down by disk and type of operation (e.g., disk /dev/nvme01 and operation WRITE) | | filesystem\_size | bytes | The size of filesystems broken down by type and device (e.g., filesystem of type ext4 for device /dev/nvme01) | | filesystem\_used | bytes | The amount of space used on the filesystems broken down by type and device (e.g., filesystem of type zfs on device /dev/nvme01) | | filesystem\_util | percent | The space utilization % of filesystems broken down by type and device (e.g., filesystem of type overlayfs on device /dev/loop1) | ### Network | Metric | | Description | | ---------------------------- | --------- | ------------------------------------------------------------------------------------------------------ | | network\_in\_bytes\_details | bytes/s | The number of inbound network packets in bytes per second broken down by network device (e.g., wlp4s0) | | network\_out\_bytes\_details | bytes/s | The number of outbound network packets in bytes per second broken down by network device (e.g., eth01) | | network\_tcp\_retrans | retrans/s | The number of network TCP retransmissions per second | ### Others | Metric | | Description | | ------------------- | ---------- | ----------------------------------------------------------------- | | os\_context\_switch | switches/s | The number of context switches per second | | proc\_blocked | processes | The number of processes blocked (e.g, for IO or swapping reasons) | ## Parameters ### CPU | Parameter | Type | Unit | Default Value | Domain | Restart | Description | | ----------------------------- | ------- | ----------- | ------------- | ------------------- | ------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | os\_cpuSchedMinGranularity | integer | nanoseconds | 1500000 | 300000 → 30000000 | no | Minimal preemption granularity (in nanoseconds) for CPU bound tasks | | os\_cpuSchedWakeupGranularity | integer | nanoseconds | 2000000 | 400000 → 40000000 | no | Scheduler Wakeup Granularity (in nanoseconds) | | os\_CPUSchedMigrationCost | integer | nanoseconds | 500000 | 100000 → 5000000 | no | Amount of time (in nanoseconds) after the last execution that a task is considered to be "cache hot" in migration decisions. A "hot" task is less likely to be migrated to another CPU, so increasing this variable reduces task migrations | | os\_CPUSchedChildRunsFirst | integer | | 0 | `0`, `1` | no | A freshly forked child runs before the parent continues execution | | os\_CPUSchedLatency | integer | nanoseconds | 12000000 | 2400000 → 240000000 | no | Targeted preemption latency (in nanoseconds) for CPU bound tasks | | os\_CPUSchedAutogroupEnabled | integer | | 0 | `0`, `1` | no | Enables the Linux task auto-grouping feature, where the kernel assigns related tasks to groups and schedules them together on CPUs to achieve higher performance for some workloads | | os\_CPUSchedNrMigrate | integer | | 32 | 3 → 320 | no | Scheduler NR Migrate | ### Memory | Parameter | Type | Unit | Default Value | Domain | Restart | Description | | ------------------------------------ | ----------- | ------------ | ------------- | ------------------------------------------------------ | ------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | os\_MemorySwappiness | integer | percent | 60 | 0 → 100 | no | The percentage of RAM free space for which the kernel will start swapping pages to disk | | os\_MemoryVmVfsCachePressure | integer | | 100 | 10 → 100 | no | VFS Cache Pressure | | os\_MemoryVmCompactionProactiveness | integer | | 20 | 0 → 100 | | Determines how aggressively compaction is done in the background | | os\_MemoryVmPageLockUnfairness | integer | | 5 | 0 → 1000 | no | Set the level of unfairness in the page lock queue. | | os\_MemoryVmWatermarkScaleFactor | integer | | 10 | 0 → 1000 | no | The amount of memory, expressed as fractions of 10'000, left in a node/system before kswapd is woken up and how much memory needs to be free before kswapd goes back to sleep | | os\_MemoryVmWatermarkBoostFactor | integer | | 15000 | 0 → 30000 | no | The level of reclaim when the memory is being fragmented, expressed as fractions of 10'000 of a zone's high watermark | | os\_MemoryVmMinFree | integer | | 67584 | 10240 → 1024000 | no | Minimum Free Memory (in kbytes) | | os\_MemoryTransparentHugepageEnabled | categorical | | `madvise` | `always`, `never`, `madvise` | no | Transparent Hugepage Enablement Flag | | os\_MemoryTransparentHugepageDefrag | categorical | | `madvise` | `always`, `never`, `defer+madvise`, `madvise`, `defer` | no | Transparent Hugepage Enablement Defrag | | os\_MemorySwap | categorical | | `swapon` | `swapon`, `swapoff` | no | Memory Swap | | os\_MemoryVmDirtyRatio | integer | | 20 | 1 → 99 | no | When the dirty memory pages exceed this percentage of the total memory, processes are forced to write dirty buffers during their time slice instead of continuing to write | | os\_MemoryVmDirtyBackgroundRatio | integer | | 10 | 1 → 99 | no | When the dirty memory pages exceed this percentage of the total memory, the kernel begins to write them asynchronously in the background | | os\_MemoryVmDirtyExpire | integer | centiseconds | 3000 | 300 → 30000 | no | When the dirty memory pages exceed this percentage of the total memory, processes are forced to write dirty buffers during their time slice instead of continuing to write | | os\_MemoryVmDirtyWriteback | integer | centiseconds | 500 | 50 → 5000 | no | Memory Dirty Writeback (in centisecs) | ### Network | Parameter | Type | Unit | Default Value | Domain | Restart | Description | | --------------------------------------- | ------- | ------------ | ------------- | --------------- | ------- | ------------------------------------------------------------------------------------------------------------------------------------------------------- | | os\_NetworkNetCoreSomaxconn | integer | megabytes | 128 | 12 → 8192 | no | Network Max Connections | | os\_NetworkNetCoreNetdevMaxBacklog | integer | megabytes/s | 1000 | 100 → 10000 | no | Network Max Backlog | | os\_NetworkNetIpv4TcpMaxSynBacklog | integer | milliseconds | 256 | 52 → 5120 | no | Network IPV4 Max Sync Backlog | | os\_NetworkNetCoreNetdevBudget | integer | | 300 | 30 → 30000 | no | Network Budget | | os\_NetworkNetCoreRmemMax | integer | | 212992 | 21299 → 2129920 | no | Maximum network receive buffer size that applications can request | | os\_NetworkNetCoreWmemMax | integer | | 212992 | 21299 → 2129920 | no | Maximum network transmit buffer size that applications can request | | os\_NetworkNetIpv4TcpSlowStartAfterIdle | integer | | 1 | `0`, `1` | no | Network Slow Start After Idle Flag | | os\_NetworkNetIpv4TcpFinTimeout | integer | | 60 | 6 → 600 | no | Network TCP timeout | | os\_NetworkRfs | integer | | 0 | 0 → 131072 | no | If enabled increases datacache hitrate by steering kernel processing of packets to the CPU where the application thread consuming the packet is running | ### Storage | Parameter | Type | Unit | Default Value | Domain | Restart | Description | | ------------------------- | ------- | --------- | ------------- | ------------------------------------- | ------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | os\_StorageReadAhead | integer | kilobytes | 128 | 0 → 4096 | no | Read-ahead speeds up file access by pre-fetching data and loading it into the page cache so that it can be available earlier in memory instead of from disk | | os\_StorageNrRequests | integer | | 32 | 12 → 1280 | no | Storage Number of Requests | | os\_StorageRqAffinity | integer | | 1 | `1`, `2` | no | Storage Requests Affinity | | os\_StorageQueueScheduler | integer | | `none` | `none`, `kyber`, `mq-deadline`, `bfq` | no | Storage Queue Scheduler Type | | os\_StorageNomerges | integer | | 0 | 0 → 2 | no | Enables the user to disable the lookup logic involved with IO merging requests in the block layer. By default (0) all merges are enabled. With 1 only simple one-hit merges will be tried. With 2 no merge algorithms will be tried | | os\_StorageMaxSectorsKb | integer | kilobytes | 256 | 32 → 256 | no | The largest IO size that the OS can issue to a block device |