The OSQuery provider

OSQuery driver documentation

The OSQuery data provider can read OSQuery log files and provide convenient query functions for each OSQuery “table” (or event type) contained in the logs.

The provider can read in one or more log files, or multiple log files in multiple folders. The files are read, converted to pandas DataFrames and grouped by table/event. In addition, date fields within the data are converted to pandas Timestamp format.

The query provider query functions will ignore parameters and do no further filtering. You can use pandas to do additional filtering and sorting of the data, or use it directly with other MSTICPy functionality.

OSQuery Configuration

You can store your connection details in msticpyconfig.yaml, instead of supplying the data_paths parameter to the QueryProvider class.

For more information on using and configuring msticpyconfig.yaml see msticpy Package Configuration and MSTICPy Settings Editor

The OSQuery settings in the file should look like the following:

DataProviders:
    ...
    OSQuery:
        data_paths:
            - /home/user1/sample_data
            - /home/shared/sample_data
        cache_file: ~/.msticpy/os_query_cache.pkl

The cache_file entry is explained later.

Expected log file format

The log file format must be a text file of JSON records. An example is shown below

{"name":"pack_osquery-snapshots-pack_python_packages","hostIdentifier":"jumpvm","calendarTime":"Thu Mar 16 09:22:33 2023 UTC","unixTime":1678958553,"epoch":0,"counter":0,"numerics":false,"decorations":{"host_uuid":"40443dd9-5b21-a345-8f89-aadde84c3719","username":"LOGIN"},"columns":{"author":"Python Packaging Authority","directory":"/usr/lib/python3.9/site-packages/","license":"UNKNOWN","name":"setuptools","path":"/usr/lib/python3.9/site-packages/setuptools-50.3.2.dist-info/","summary":"Easily download, build, install, upgrade, and uninstall Python packages","version":"50.3.2"},"action":"snapshot"}
{"name":"pack_osquery-snapshots-pack_dns_resolvers","hostIdentifier":"jumpvm","calendarTime":"Thu Mar 16 13:14:10 2023 UTC","unixTime":1678972450,"epoch":0,"counter":0,"numerics":false,"decorations":{"host_uuid":"40443dd9-5b21-a345-8f89-aadde84c3719","username":"LOGIN"},"columns":{"address":"168.63.129.16","id":"0","netmask":"32","options":"705","type":"nameserver"},"action":"snapshot"}

Each JSON record is expected to have a name field, identifying the event type, along with child dictionaries (columns and decorations.

Using the OSQuery provider

To use the OSQuery provider you need to create an QueryProvider instance, passing the string “OSQueryLogs” as the data_environment parameter. If you have not configured data_paths in msticpyconfig.yaml, you also need to add the data_paths parameter to specify specific folders or files that you want to read.

Calling the connect method triggers the provider to read the log files.

100%|██████████| 2/2 [00:00<00:00, 25.01it/s]
Data loaded.

Listing OSQuery tables

qry_prov.list_queries()

['osquery.acpi_tables',
'osquery.device_nodes',
'osquery.dns_resolvers',
'osquery.events',
'osquery.fim',
'osquery.last',
'osquery.listening_ports',
'osquery.logged_in_users',
'osquery.mounts',
'osquery.open_sockets',
'osquery.osquery_info',
'osquery.osquery_packs',
'osquery.osquerydb_size',
'osquery.platform_info',
'osquery.process_memory',
'osquery.processes',
'osquery.python_packages',
'osquery.schedule',
'osquery.shell_history']

Running an OSQuery query

Each query returns the table of event types retrieved from the logs.

qry_prov.osquery.processes()

name	hostIdentifier	unixTime	…	username	tname_	parent	username
pack_osquery-custom-pack_processes	jumpvm	2023-03-16 03:08:58+00:00	…	LOGIN	kthreadd	2	root
pack_osquery-custom-pack_processes	jumpvm	2023-03-16 03:08:58+00:00	…	LOGIN	kthreadd	2	root
pack_osquery-custom-pack_processes	jumpvm	2023-03-16 03:08:58+00:00	…	LOGIN	kthreadd	2	root
pack_osquery-custom-pack_processes	jumpvm	2023-03-16 03:08:58+00:00	…	LOGIN	kthreadd	2	root
pack_osquery-custom-pack_processes	jumpvm	2023-03-16 03:08:58+00:00	…	LOGIN	kthreadd	2	root

Note

Columns in the the nested log data may be renamed if their name clashes with an existing name. See the example name_ in the previous table.