DataKit AIX

DataKit AIX collects IBM AIX host metrics, host objects, errpt error logs, and can optionally enable middleware metrics for DB2, Oracle, Informix, Sybase, IBM MQ, WebSphere, IHS, SAP, and other products.

Tested versions:

AIX 7.2 TL5 SP10

Configuration¶

Installation¶

Run the installation script on the AIX host. DATAKITAIX_DATAWAY is written to the [dataway] section in /etc/datakitaix/datakitaix.toml.

Install the latest version:

DATAKITAIX_DATAWAY=https://openway.guance.com?token=<TOKEN> sh -c "$(curl -L https://static.guance.com/datakitaix/install.sh)"

Install a specific version:

DATAKITAIX_DATAWAY=https://openway.guance.com?token=<TOKEN> sh -c "$(curl -L https://static.guance.com/datakitaix/install-<VERSION>.sh)"

Example:

DATAKITAIX_DATAWAY=https://openway.guance.com?token=<TOKEN> sh -c "$(curl -L https://static.guance.com/datakitaix/install-1.0.0.sh)"

Default paths after installation:

Path	Description
`/opt/datakitaix/bin/datakitaix`	Main program
`/usr/bin/datakitaix`	Command symlink for running `datakitaix` directly
`/etc/datakitaix/datakitaix.toml`	Runtime configuration generated during installation
`/etc/datakitaix/datakitaix.toml.sample`	Configuration sample for the current version; updated during upgrade
`/var/lib/datakitaix`	Runtime data, spool, and errpt cursor
`/var/log/datakitaix`	Runtime logs

The installation script registers the AIX SRC service datakitaix. If DataWay is not configured, the service is registered but not started automatically.

Upgrade¶

Run the following command on a host where DataKit AIX is already installed:

DATAKITAIX_UPGRADE=1 sh -c "$(curl -L https://static.guance.com/datakitaix/install.sh)"

To update DataWay during upgrade, also set DATAKITAIX_DATAWAY:

DATAKITAIX_DATAWAY=https://openway.guance.com?token=<TOKEN> DATAKITAIX_UPGRADE=1 sh -c "$(curl -L https://static.guance.com/datakitaix/install.sh)"

The upgrade replaces the binary, updates datakitaix.toml.sample, and preserves the existing /etc/datakitaix/datakitaix.toml. If DATAKITAIX_DATAWAY rewrites an existing configuration during upgrade, datakitaix.toml.old.<timestamp> is created in the same directory first.

Uninstall¶

/opt/datakitaix/bin/uninstall-aix.sh

By default, uninstall removes only the binary, helper script, command symlink, and SRC registration, while preserving configuration and runtime data. To also remove configuration and data:

PURGE=1 /opt/datakitaix/bin/uninstall-aix.sh

Commands¶

Command	Description
`datakitaix run -config /etc/datakitaix/datakitaix.toml`	Run in the foreground and start collectors, inputs, log tailers, health checks, and OpenTelemetry according to the configuration.
`datakitaix once -config /etc/datakitaix/datakitaix.toml`	Run collectors once and report data to DataWay.
`datakitaix once -dry-run -config /etc/datakitaix/datakitaix.toml`	Run collectors once and print line protocol without reporting data.
`datakitaix check -config /etc/datakitaix/datakitaix.toml`	Validate configuration and run startup checks.
`datakitaix version`	Print the current version.
`startsrc -s datakitaix`	Start the AIX SRC service.
`stopsrc -s datakitaix`	Stop the AIX SRC service.
`lssrc -s datakitaix`	View AIX SRC service status.

Main configuration¶

The main configuration keeps only global settings. See the integration documents below for host collectors, middleware inputs, file logs, health checks, and OpenTelemetry configuration.

[logging]
# Agent runtime log configuration.
log = "/var/log/datakitaix/datakitaix.log"
error_log = "/var/log/datakitaix/error.log"
level = "info"
disable_color = false
rotate = 32
rotate_backups = 5

[dataway]
# DataWay URL. Setting `DATAKITAIX_DATAWAY` during installation writes this configuration automatically.
urls = ["https://openway.guance.com?token=<TOKEN>"]
timeout = "30s"
retry = 2
gzip = true
max_raw_body_size = 1048576

[spool]
# Enable local disk buffering for retryable upload failures.
enabled = true
# Directory for failed data buffering; also stores log tailer offset state.
path = "/var/lib/datakitaix/spool"
# Maximum disk usage for pending spool files. When exceeded, the oldest files are dropped first.
max_size_mb = 128
# Maximum number of pending spool files.
max_files = 5000
# Retry interval for pending spool data.
drain_interval = "30s"

[global_tags]
# Global tags appended to all collected data.
# env = "prod"

[self_monitor]
# Report datakitaix process, Go runtime, and spool status.
enabled = true
# Self-monitoring metric reporting interval.
interval = "30s"

Host collectors¶

Host collectors collect AIX host metrics, process objects, host objects, and errpt data.

[collectors]
enabled = true
command_timeout = "10s"

Collector Name	Description
CPU (`cpu`)	Collect host CPU usage.
Disk (`disk`)	Collect disk and filesystem usage.
Disk I/O (`diskio`)	Collect host disk I/O metrics.
Mem (`mem`)	Collect host memory usage.
Swap (`swap`)	Collect paging space usage.
System (`system`)	Collect host operating system load.
Net (`net`)	Collect host network traffic and protocol stack metrics.
LPARStat (`lparstat`)	Collect LPAR and hypervisor metrics. Disabled by default.
Process (`process`)	Collect host process objects.
Host Object (`hostobject`)	Collect basic host information, such as operating system and hardware information.
Errpt (`errpt`)	Collect AIX error reports.

Middleware Inputs¶

Middleware inputs are disabled by default. Set [inputs].enabled = true and enable the specific instance to collect them.

[inputs]
enabled = true
command_timeout = "15s"

Collector Name	Description
DB2 (`db2`)	Collect IBM DB2 instance, database, tablespace, and HADR status.
Oracle (`oracle`)	Collect Oracle database, tablespace, ASM, FRA, and wait event metrics.
Informix (`informix`)	Collect Informix instance, session, dbspace, and log usage.
Sybase (`sybase`)	Collect Sybase ASE connections and database capacity.
IBM MQ (`ibm_mq`)	Collect Queue Manager, queue, channel, and listener status.
WebSphere (`websphere`)	Collect WebSphere server, JVM, thread pool, and HTTP session metrics.
IHS (`ihs`)	Collect IBM HTTP Server worker, request count, and scoreboard metrics.
SAP (`sap`)	Collect SAP instance processes, queue, enqueue, and alert metrics.

File Logs¶

File Logs Local file logs are collected with [logs] and one or more [[logs.files]] blocks.

Health Check¶

health-check TCP/HTTP availability checks are configured with [healthcheck].

OpenTelemetry¶

OpenTelemetry The OpenTelemetry HTTP/gRPC receiver is configured with [otel].

Self Metrics¶

DataKit AIX can report self-monitoring metrics through [self_monitor] to observe the datakitaix process, Go runtime, and spool status.

`datakitaix`¶

Tags & Fields	Description
host (`tag`)	Hostname.
uptime_seconds	datakitaix process uptime. Type: int \| (gauge) Unit: s
goroutines	Current number of Go goroutines. Type: int \| (gauge) Unit: count
cpu_usage	CPU usage of the datakitaix process. Type: float \| (gauge) Unit: percent
cpu_cores	Number of CPU cores visible to the Go runtime. Type: int \| (gauge) Unit: count
go_mem_heap_alloc	Current bytes allocated on the Go heap. Type: int \| (gauge) Unit: byte
go_mem_heap_inuse	Current bytes in use on the Go heap. Type: int \| (gauge) Unit: byte
go_mem_heap_released	Bytes returned by the Go heap to the operating system. Type: int \| (gauge) Unit: byte
go_mem_stack_inuse	Current bytes used by Go stacks. Type: int \| (gauge) Unit: byte
go_mem_sys	Total bytes requested by the Go runtime from the operating system. Type: int \| (gauge) Unit: byte
go_gc_count	Cumulative number of Go GC cycles. Type: int \| (count) Unit: count
go_gc_pause_total_ns	Cumulative Go GC pause time. Type: int \| (count) Unit: ns
spool_pending	Current number of spool files pending send. Type: int \| (gauge) Unit: count
spool_bytes	Current number of spool bytes pending send. Type: int \| (gauge) Unit: byte
spool_bad_files	Current number of damaged spool files. Type: int \| (gauge) Unit: count
spool_bad_bytes	Current bytes in damaged spool files. Type: int \| (gauge) Unit: byte
spool_tmp_files	Current number of temporary spool files. Type: int \| (gauge) Unit: count
spool_tmp_bytes	Current bytes in temporary spool files. Type: int \| (gauge) Unit: byte
spool_dropped_files	Current number of dropped spool files. Type: int \| (gauge) Unit: count
spool_dropped_bytes	Current bytes in dropped spool files. Type: int \| (gauge) Unit: byte

For OpenTelemetry receiver self metrics datakitaix_otel_receiver, see OpenTelemetry.