Awesome

go-audit

About

go-audit is an alternative to the auditd daemon that ships with many distros. After having created an auditd audisp plugin to convert audit logs to json, I became interested in creating a replacement for the existing daemon.

Goals

Safe : Written in a modern language that is type safe and performant
Fast : Never ever ever ever block if we can avoid it
Outputs json : Yay
Pluggable pipelines : Can write to syslog, local file, or stdout. Additional outputs are easily written.
Connects to the linux kernel via netlink (info here and here)

Usage

Installation

Install golang, version 1.7 or greater is required
Install govendor if you haven't already

go get -u github.com/kardianos/govendor
Clone the repo
```
git clone (this repo)
cd go-audit
```
Build the binary
```
make
```
Copy the binary go-audit to wherever you'd like

Testing

make test - run the unit test suite
make test-cov-html - run the unit tests and open up the code coverage results
make bench - run the benchmark test suite
make bench-cpu - run the benchmark test suite with cpu profiling
make bench-cpulong - run the benchmark test suite with cpu profiling and try to get some gc collection

Running as a service

Check the contrib folder, it contains examples for how to run go-audit as a proper service on your machine.

Example Config

See go-audit.yaml.example

FAQ

I am seeing `Error during message receive: no buffer space available` in the logs

This is because go-audit is not receiving data as quickly as your system is generating it. You can increase the receive buffer system wide and maybe it will help. Best to try and reduce the amount of data go-audit has to handle.

If reducing audit velocity is not an option you can try increasing socket_buffer.receive in your config. See Example Config for more information

socket_buffer:
    receive: <some number bigger than (the current value * 2)>

Sometime files don't have a `name`, only `inode`, what gives?

The kernel doesn't always know the filename for file access. Figuring out the filename from an inode is expensive and error prone.

You can map back to a filename, possibly not the filename, that triggured the audit line though.

sudo debugfs -R "ncheck <inode to map>" /dev/<your block device here>

I don't like math and want you to tell me the syslog priority to use

Use the default, or consult this handy table.

Wikipedia has a pretty good page on this

	emerg (0)	alert (1)	crit (2)	err (3)	warn (4)	notice (5)	info (6)	debug (7)
kernel (0)	0	1	2	3	4	5	6	7
user (1)	8	9	10	11	12	13	14	15
mail (2)	16	17	18	19	20	21	22	23
daemon (3)	24	25	26	27	28	29	30	31
auth (4)	32	33	34	35	36	37	38	39
syslog (5)	40	41	42	43	44	45	46	47
lpr (6)	48	49	50	51	52	53	54	55
news (7)	56	57	58	59	60	61	62	63
uucp (8)	64	65	66	67	68	69	70	71
clock (9)	72	73	74	75	76	77	78	79
authpriv (10)	80	81	82	83	84	85	86	87
ftp (11)	88	89	90	91	92	93	94	95
ntp (12)	96	97	98	99	100	101	102	103
logaudit (13)	104	105	106	107	108	109	110	111
logalert (14)	112	113	114	115	116	117	118	119
cron (15)	120	121	122	123	124	125	126	127
local0 (16)	128	129	130	131	132	133	134	135
local1 (17)	136	137	138	139	140	141	142	143
local2 (18)	144	145	146	147	148	149	150	151
local3 (19)	152	153	154	155	156	157	158	159
local4 (20)	160	161	162	163	164	165	166	167
local5 (21)	168	169	170	171	172	173	174	175
local6 (22)	176	177	178	179	180	181	182	183
local7 (23)	184	185	186	187	188	189	190	191

I am seeing duplicate entries in syslog!

This is likely because you are running journald which is also reading audit events. To disable it you need to disable the functionality in journald.

sudo systemctl mask systemd-journald-audit.socket

Thanks!

To Hardik Juneja, Arun Sori, Aalekh Nigam Aalekhn for the inspiration via https://github.com/mozilla/audit-go