This is gcp-flowlogs-reader, a command line tool and Python library for
retrieving and manipulating VPC Flow Logs for the Google Cloud platform.
VPC Flow Logs record metadata about network communication inside your Google Cloud VPC. They can be used for security monitoring and performance analysis, and analogous to NetFlow and IPFIX records for on-premises networks.
For more information about VPC Flow Logs in Google Cloud, see these docs.
Note that this branch wraps version 1 of the Python Client for Cloud Logging.
Clone this repository and run python setup.py install to install the
library.
Once it's installed, the gcp_flowlogs_reader tool should be available to run.
You may also invoke it with python -m gcp_flowlogs_reader.
$ gcp_flowlogs_reader -h
usage: gcp_flowlogs_reader [-h] [--start-time START_TIME]
[--end-time END_TIME] [--time-format TIME_FORMAT]
[--filters FILTERS]
[--credentials-file CREDENTIALS_FILE]
[--log-name LOG_NAME]
[action [action ...]]
If the GOOGLE_APPLICATION_CREDENTIALS environment variable is pointing to
a JSON file holding service account credentials, the tool will use that.
Otherwise, authenticate per
Google instructions
or set the --credentials-file switch.
By default, the tool will print VPC Flow Log records from your project under the default name. It will look back one hour:
$ gcp_flowlogs_reader --credentials-file="/home/service-account-.json"
src_ip dest_ip src_port dest_port protocol start_time end_time bytes_sent packets_sent
192.0.2.2 198.51.100.53 22 53658 6 2018-04-03T19:07:59 2018-04-03T19:08:00 3220 12
203.0.113.75 192.0.2.2 57772 22 6 2018-04-03T19:08:11 2018-04-03T19:08:12 5400 13
198.51.100.53 192.0.2.2 53658 22 6 2018-04-03T19:08:00 2018-04-03T19:08:00 2836 4
192.0.2.2 203.0.113.75 3389 57772 6 2018-04-03T19:08:11 2018-04-03T19:08:11 3384 8You can also specify gcp_flowlogs_reader print <limit> to limit the number
printed.
To change the default one-hour lookback, set --start-time or --end-time.
The default format is %Y-%m-%d %H:%M:%S and assumes the UTC time zone.
You can change it with the --time-format switch.
See strftime(3) or the Python documentation for
strptime
for information on constructing format strings.
To print the set of internal and external IP addresses in the flow records,
use the ipset action:
$ gcp_flowlogs_reader --start-time="2018-04-03 14:00:00" ipset
192.0.2.2
192.0.2.3
198.51.100.53
203.0.113.75To find all flows associated with a single IP address, use the findip action:
$ python -m gcp_flowlogs_reader findip 198.51.100.3
src_ip dest_ip src_port dest_port protocol start_time end_time bytes_sent packets_sent
198.51.100.3 192.0.2.3 34993 3389 6 2018-04-03T19:21:40 2018-04-03T19:21:40 0 9
198.51.100.3 192.0.2.3 34993 3389 6 2018-04-03T19:21:40 2018-04-03T19:21:40 459 11
198.51.100.3 192.0.2.3 60504 3389 6 2018-04-03T19:22:32 2018-04-03T19:22:32 258 6
192.0.2.3 198.51.100.3 3389 60504 6 2018-04-03T19:22:32 2018-04-03T19:22:34 3256 10To aggregate flows with the same 5-tuple key (source address, source port,
destination address, destination port, protocol), use the aggregate action.
gcp_flowlogs_reader.FlowRecord transforms the JSON Payload from the
VPC Flow Logs entry into a Python object with standard types:
| Attribute | Type | Example |
|---------------|----------------------|------------------------------------------------------------------------------------------------------|
| src_ip | ipaddress.ip_address | ipaddress.ip_address('192.0.2.2') |
| src_port | int | 3389 |
| dest_ip | ipaddress.ip_address | ipaddress.ip_address('198.51.100.1') |
| dest_port | int | 49152 |
| protocol | int | 6 |
| start_time | datetime.datetime | datetime.datetime(2018, 4, 4, 11, 55, 12, 943517) |
| end_time | datetime.datetime | datetime.datetime(2018, 4, 4, 11, 55, 14, 125304) |
| bytes_sent | int | 4446 |
| packets_sent | int | 13 |
| rtt_msec | int | 233 |
| reporter | str | 'SRC' |
| src_instance | namedtuple | InstanceDetails(project_id='yoyodyne-1020', vm_name='vm-1020', region='us-west1', zone='us-west1-a') |
| dest_instance | namedtuple | InstanceDetails(project_id='yoyodyne-1020', vm_name='vm-1020', region='us-west1', zone='us-west1-a') |
| src_vpc | namedtuple | VpcDetails(project_id='yoyo-1020', vpc_name='prod-vpc-3', subnetwork_name='prod-net-3') |
| dest_vpc | namedtuple | VpcDetails(project_id='yoyo-1020', vpc_name='prod-vpc-3', subnetwork_name='prod-net-3') |
| src_location | namedtuple | GeographicDetails(continent='America', country='usa', region='California', city='Santa Teresa') |
| dest_location | namedtuple | GeographicDetails(continent='America', country='usa', region='California', city='Santa Teresa') |
gcp_flowlogs_reader.Reader acts as an iterator over flows from the logs:
from gcp_flowlogs_reader import Reader
credentials_path = '/home/service-account-.json'
ip_set = set()
for flow_record in Reader(service_account_json=credentials_path):
ip_set.add(flow_record.src_ip)
ip_set.add(flow_record.dest_ip)
print(len(ip_set))You may pass in these keywords as arguments to affect what flows are returned:
log_name- defaults toprojects/{#project_id}/logs/compute.googleapis.com%2Fvpc_flowsstart_time- defaults to one hour agoend_time- defaults to nowfilters- defaults the the standard log namecollect_multiple_projects- defaults to Falselogging_client- a customgoogle.cloud.logging.Clientinstanceservice_account_json- the path to a service account JSON credential fileservice_account_info- the service account information parsed from a credential file. Seefrom_service_accountfor more information.