You can download the audit trail of a session from the /api/audit/sessions/<session-id>/audit_trail endpoint. To find a specific audit trail, see Searching in the session database. You can download audit trails that are available on SPS, and also audit trails that have been archived (if SPS can access the archived audit trail).
curl --cookie cookies "https://<IP-address-of-SPS>/api/audit/sessions/<session-id>/audit_trail"
To actually create a file, you must save the downloaded data into a file (use the .zat file extension), for example:
curl --cookie cookies "https://<IP-address-of-SPS>/api/audit/sessions/<session-id>/audit_trail" > my-downloaded-trail.zat
You can replay the downloaded audit trails with the Safeguard Desktop Player application. For details, see Safeguard Desktop Player User Guide.
If you want to replay an ongoing session in follow mode, you have to download the audit trail in .srs format. Use the ?format=srs option:
curl --cookie cookies "https://<IP-address-of-SPS>/api/audit/sessions/<session-id>/audit_trail?format=srs" > my-downloaded-trail.srs
For details, see Replaying audit files in follow mode in the Safeguard Desktop Player User Guide.
You can list, search, and filter the SPS session database at the /api/audit/sessions endpoint. You can use the following actions:
-
?start
Display sessions that started after the specified date. Use the ISO 8601 format for the date, for example, 2017-01-25T10:00.
-
?end
Display sessions that ended before the specified date. Use the ISO 8601 format for the date, for example, 2017-01-25T10:00.
-
?fields
Display the selected properties (elements and values) of the listed sessions.
-
?q
Filter the list using one or more properties (elements) of the sessions.
-
?limit
The ?limit parameter allows you to configure the maximum number of results to display on a page at once.
The default value of ?limit is 500.
NOTE: The default value of 500 is the maximum permitted value you can set for ?limit. If you set the ?limit parameter to a value bigger than 500, only the first 500 results will be displayed.
-
?offset
The ?offset parameter allows you to configure the offset from the first result that is displayed. This can be useful if the number of items returned exceeds the number of items displayed on the first page, and you want to navigate to any of the subsequent items displayed on other pages.
The default value of ?offset is null.
NOTE: The maximum number of search results in One Identity Safeguard for Privileged Sessions is 10000. As a result, any ?offset values set to larger than 10000 will be ignored and the results exceeding the value of 10000 will not be displayed.
-
?sort
Sort the results based on the values of the fields.
-
?format
Configure the format of the displayed results.
The default value of ?format is json. If you do not configure the ?format parameter, the results will be displayed in JSON format.
To display search results in a CSV format, enter csv as a value.
To combine multiple expressions, use the & (ampersand) character, for example:
Display the target server and port of each active session:
curl --cookie cookies "https://<IP-address-of-SPS>/api/audit/sessions?fields=psm.target.address,psm.target.port&q=active:true"
Display 10 sessions at once, and navigate to 31-40:
curl --cookie cookies "https://<IP-address-of-SPS>/api/audit/sessions?limit=10&offset=31"
Search in metadata and session content at the same time:
curl --cookie cookies "https://<IP-address-of-SPS>/api/audit/sessions?q=protocol:ssh&content=sudo"
NOTE: If you use curl, use quotation marks for the URL to avoid problems with the & (ampersand) character.
Response
The response to search or filtering action contains a list of the matching sessions, as well as some additional meta fields. For example:
{
"items": [
{
"body": {
"duration": 0,
"name": "myname",
"start_time": "2017-01-25T11:11:52.000+01:00"
},
"key": "2",
"meta": {
"href": "/api/audit/sessions/2"
}
},
{
"body": {
"duration": 34,
"name": "myname",
"start_time": "2017-01-25T11:11:11.000+01:00"
},
"key": "10",
"meta": {
"href": "/api/audit/sessions/10"
}
}
],
"meta": {
"fields": [
"start_time",
"name",
"duration"
],
"first": "/api/audit/sessions?limit=500&offset=0&fields=start_time,name,duration&q=name%3Amyname&=duration",
"href": "/api/audit/sessions",
"last": "/api/audit/sessions?limit=500&offset=0&fields=start_time,name,duration&q=name%3Amyname&sort=duration",
"limit": 500,
"match_count": 2,
"next": null,
"offset": 0,
"parent": "/api/audit",
"previous": null
}
items |
|
list |
Top level element, a list containing the details of the matching sessions. |
|
body |
JSON object |
Contains the information returned about a session, that is, the fields selected with the ?fields expression. For example, if you used the fields=start_time,psm.gateway_username,duration expression in your query, then the body element contains these fields for each returned session: "body": {
"duration": 0,
"name": null,
"start_time": "2017-01-25T11:11:52.000+01:00"
}, |
|
key |
string |
A globally unique string that identifies the session. This session ID has the following format: svc/<unique-random-hash>/<name-of-the-connection-policy>:<session-number-since-service-started>/<protocol>, for example, svc/5tmEaM7xdNi1oscgVWpbZx/ssh_console:1/ssh.
Log messages related to the session also contain this ID. For example: 2015-03-20T14:29:15+01:00 demo.example
zorp/scb_ssh[5594]: scb.audit(4):
(svc/5tmEaM7xdNi1oscgVWpbZx/ssh_console:0/ssh):
Closing connection; connection='ssh_console',
protocol='ssh', connection_id='409829754550c1c7a27e7d',
src_ip='10.40.0.28', src_port='39183',
server_ip='10.10.20.35', server_port='22',
gateway_username='', remote_username='example-username',
verdict='ZV_ACCEPT'
Note that when using the session ID in a REST call, you must replace the special characters in the ID with the hyphen (-) character. For example, if the session ID in the log message is svc/fNLgRmAyf5EtycgUYnKc1B/ssh_demo2:2, use the svc-fNLgRmAyf5EtycgUYnKc1B-ssh_demo2-2 ID in REST calls. |
In addition to the usual meta elements of other endpoints, search results can contain the following additional elements.
meta |
|
JSON object |
Top level element, a list containing meta information about the response. |
|
fields |
list |
Contains the list of data fields returned about each session, that is, the fields selected with the ?fields expression. For example, if you used the fields=start_time,psm.gateway_username,duration expression in your query, then the body element contains these fields for each returned session: "fields": [
"start_time",
"name",
"duration"
], |
|
limit |
integer |
The maximum number of sessions returned in a the response (by default, 500). |
|
match_count |
integer |
The number of results matching the query. |
|
next |
string |
A query to retrieve the next set of search results, if match_count is higher than limit. |
|
offset |
integer |
Indicates the position of the results in this response, relative to the total number of results (match_count). Otherwise, its value is null. |
|
previous |
string |
A query to retrieve the previous set of search results, if match_count is higher than limit, and offset is higher than 0. Otherwise, its value is null. |
Filtering
You can use the ?q option to filter the list using one or more property (element) of the sessions.
?q=protocol:ssh
You can escape special characters using the backslash character.
?q=server_username:\"Windows User\"
To add multiple elements to the filter, you can use the AND, AND NOT, and OR operators.
?q=protocol:ssh AND verdict:accept AND NOT name:admin
You can create groups using () (parentheses).
?q=(client.address:10.20.30.40 OR target.address:10.20.30.40) AND verdict:accept
You can also use () (parentheses) to add multiple possible values for a property.
?q=protocol:(ssh rdp)
You can use the * (asterisk) and ? (question mark) wildcards for string-type values.
?q=name:?dmi*
You can define ranges using [] (brackets) or {} (braces) and the TO operator. This only works for numeric (int) values.
-
[ means equal or higher than the following value
-
] means equal or lower than the preceding value
-
{ means higher than the following value
-
}means lower than the preceding value
For example, the following range resolves to 22:
?q=port:{21 TO 23}
You can also use the * (asterisk) wildcard in the range.
?q=start_time:[* TO 1461654799]
Note that not all connection data can be used for filtering. The available elements are:
-
active
Boolean, true means the session is ongoing (it is still active).
-
auth_method
String, the authentication method used.
-
channel_policy
String, the key of the channel policy.
-
client.address
String, the IP address of the client.
-
client.port
Integer, the port of the client.
-
psm.connection_policy
String, the key of the connection policy.
-
end_time
The date of the end of the session in ISO 8601 format.
-
name
String, the username used for authenticating against the gateway.
-
protocol
String, the protocol of the session.
-
server.address
String, the IP of the remote server.
-
psm.server_local.address
String, the IP of SPS.
-
psm.server_local.port
String, the port of SPS.
-
server.port
String, the port of the remote server.
-
server_username
String, the username used for authenticating on the remote server.
-
session_id
String, the identifier of the session.
-
start_time
The date of the start of the session in ISO 8601 format.
-
target.address
String, the IP the client targeted in the session.
-
target.port
Integer, the port the client targeted in the session.
-
verdict
String, the connection verdict. Possible values are:
-
accept
The connection attempt was successful.
-
accept-terminated
The connection violated a content policy, and was terminated by SPS.
-
auth-fail
Authentication failure.
-
deny
The connection was denied.
-
fail
The connection attempt failed.
-
gw-auth-fail
Gateway authentication failure.
-
key-error
The connection attempt failed due to a host key mismatch.
-
user-mapping-fail
The connection attempt failed due to a user mapping failure.
Content search in indexed audit trails
You can use the ?q=screen.content option to search for keywords that appear in the screen content of the audit trails. Such content is any text that appeared on the screen in terminal or graphical sessions, or commands that the user entered in terminal sessions. Note that content search works only if:
-
Indexing was enabled in the connection policy related to the audit trail during the session, and
-
the audit trail has already been indexed.
?q=screen.content:"my-search-expression"
You can use the Apache Lucene query syntax to create the search expression, but note the following points.
-
You must format the search expression as an URL, and escape special characters accordingly. For example, if your search expression is man iptables, you must escape the whitespace: man%20iptables
For a list of special (reserved) URL characters, see RFC3986.
-
Do not begin the expression with the * wildcard.
Examples:
Search for the word example
?q=screen.content:example
Search for the words example, examples, and so on:
?q=screen.content:example%3F
Search for the words example, examine, and so on:
?q=screen.content:exam%2A
Search in metadata and session content at the same time:
curl --cookie cookies "https://<IP-address-of-SPS>/api/audit/sessions?q=protocol:ssh&content=sudo"
For further details and examples, see Searching in the contents of audit trails in the Administration Guide.
Displaying session data
You can use the ?fields option to display the selected data (body elements) of each session.
?fields=protocol
To list multiple elements, use the , (comma) character. Note that the response includes the selected fields in alphabetic order, not in the order they were specified.
?fields=protocol,name
To list all possible elements, use the fields=* expression.
?fields=*
Note that not all connection data can be displayed in the generated list. The available elements are:
-
active
Boolean, true means the connection is ongoing.
-
archived
Boolean, true means the session has been archived.
-
auth_method
String, the authentication method used.
-
channel_policy
String, the key of the channel policy.
-
client.address
String, the IP address of the client.
-
client.port
Integer, the port of the client.
-
connection_policy
String, the key of the connection policy.
-
duration
Integer, the duration of the session. Computed value.
-
end_time
The date of the end of the session in ISO 8601 format.
-
name
String, the username used for authenticating against the gateway.
-
protocol
String, the protocol of the session.
-
server.address
String, the IP of the remote server.
-
server_local.address
String, the IP of SPS.
-
server_local.port
Integer, the port of SPS.
-
server.port
Integer, the port of the remote server.
-
server_username
String, the username used for authenticating on the remote server.
-
session_id
String, the identifier of the session.
-
start_time
The date of the start of the session in ISO 8601 format.
-
target.address
String, the IP the client targeted in the session.
-
target.port
Integer, the port the client targeted in the session.
Date-specific search
To display search results only for specific date intervals, you can use the ?start and ?end options.
-
The ?start option selects the sessions that started after the specified date (based on the value of the start_time field).
-
The ?end option selects the sessions that ended before the specified date (based on the value of the end_time field).
-
Both options accept the date in ISO 8601 format.
?start=2017-01-25T11:11:52.000+01:00
?end=2017-01-25T11:41:52.000+01:00
?start=2017-01-24&end=2017-01-25
Examples:
Select sessions that started on January 20, 2017, or later:
?start=2017-01-20
Select sessions that started on 11:00 January 20, 2017, or later:
?start=2017-01-20T11:00
Select sessions that ended on January 20, 2017:
?end=2017-01-20
Select sessions started and ended on January 20, 2017:
?start=2017-01-20&end=2017-01-20
Select sessions started after 11:00, January 20, 2017, and ended before 09:00, January 21, 2017:
?start=2017-01-20T11:00&end=2017-01-21T09:00
Sort the results
You can sort the search results using the sort expression, for example, based on the length of the sessions:
?sort=duration
You can use any field to sort the results. By default, sorting returns the results in ascending order, if you use ?sort=duration, then the shortest session is at the beginning of the list. To sort the results in descending order, add the minus sign (-) before the field name. For example, the response to the following expression starts with the longest session:
?sort=-duration
You can specify multiple fields to order the list. In this case, the list is first ordered using the first field, then the second, and so on. For example, to order the list first by duration, then by start time, use the following expression.
?sort=duration,start_time
The following example sorts the results by duration, and displays the start time, gateway username, and duration fields.
curl --cookie cookies "https://<IP-address-of-SPS>/api/audit/sessions?sort=duration&fields=start_time,psm.gateway_username,duration"
Configure the format of the displayed results
The default value of ?format is json. If you do not configure the ?format parameter, the results will be displayed in JSON format.
?format=json
To display search results in a CSV format, enter csv as a value.
?format=csv
Example: querying sessions in CSV result format
Given that the following sessions were recorded:
{
"1": {
"channel": [
{"channel_id": 1},
{"channel_id": 2}
],
"recording": {
"session_id": 1,
"archived": false,
"channel_policy": "policy1",
"content_reference_id": 1,
"connection_policy": "connection1",
"auth_method": "password",
"target": {
"port": 2222,
"ip": "1.1.1.1",
"name": "1.1.1.1"
},
"server_local": {
"port": 46,
"ip": "1.1.1.1",
"name": "1.1.1.1"
}
},
"user": {
"server_username": "user1",
"gateway_username": "user1"
},
"client": {
"port": 48679,
"ip": "2.2.2.2",
"name": "2.2.2.2"
},
"active": false,
"start_time": 1,
"duration": 4,
"server": {
"port": 22,
"ip": "2.2.2.2",
"name": "2.2.2.2"
},
"end_time": 5,
"protocol": "ssh"
},
"2": {
"channel": [
{"channel_id": 3},
{"channel_id": 4}
],
"recording": {
"session_id": 2,
"archived": false,
"channel_policy": "policy2",
"content_reference_id": 2,
"connection_policy": "connection2",
"auth_method": "password",
"target": {
"port": 2222,
"ip": "1.1.1.1",
"name": "1.1.1.1"
},
"server_local": {
"port": 46,
"ip": "1.1.1.1",
"name": "1.1.1.1"
}
},
"user": {
"server_username": "user2",
"gateway_username": "user2"
},
"client": {
"port": 48680,
"ip": "3.3.3.3",
"name": "3.3.3.3"
},
"active": false,
"start_time": 1,
"duration": 4,
"server": {
"port": 24,
"ip": "2.2.2.2",
"name": "2.2.2.2"
},
"end_time": 7,
"protocol": "ssh"
}
}
When the query is the following:
curl --cookie cookies "https://<IP-address-of-SPS>/api/audit/sessions?format=csv&fields=protocol,end_time,user.gateway_username,server.ip,client.ip,client.port"
The response is the following:
"Key","Protocol","End time","Gateway username","Server IP","Client IP","Client port"
"2","ssh","7","user2","2.2.2.2","3.3.3.3","48680"
"1","ssh","5","user1","2.2.2.2","2.2.2.2","48679"
Example: querying sessions in CSV result format with interesting events
Given that the following sessions were recorded:
{
"1":{
"origin": "RECORDING",
"protocol": "SSH",
"analytics": {
"interesting_events": ["ssh", "sudo"],
"similar_sessions": []
},
"recording": {
"session_id": "1",
"verdict": "ACCEPT",
"audit_trail": "/var/lib/zorp/audit/532078660569910c6542b2/01/audit-scb_ssh-1451900800-1.zat",
"connection_policy": "ssh1",
"content_reference_id": 1
}
},
"2":{
"origin": "RECORDING",
"protocol": "SSH",
"analytics": {
"interesting_events": ["sudo", "systemctl"],
"similar_sessions": []
},
"recording": {
"session_id": "2",
"verdict": "ACCEPT",
"connection_policy": "ssh2",
"content_reference_id": 2
}
}
}
When the query is the following:
curl --cookie cookies "https://<IP-address-of-SPS>/api/audit/sessions?sort=recording.session_id&format=csv&fields=recording.session_id,analytics.interesting_events,analytics.similar_sessions"
The response is the following:
"Key","Recording Session ID","Analytics Interesting events","Similar Sessions"
"1","1","ssh",""
"1","1","sudo",""
"2","2","sudo",""
"2","2","systemctl",""
Example: querying sessions in CSV result format with audit trail link
Given that the following sessions were recorded:
{
"svc-paKzcMJwXghEFJ9UvsdqFU-sid-1": {
"origin": "RECORDING",
"protocol": "SSH",
"recording": {
"session_id": "1",
"verdict": "ACCEPT",
"audit_trail": "/var/lib/zorp/audit/532078660569910c6542b2/01/audit-scb_ssh-1451900800-1.zat",
"connection_policy": "ssh1",
"content_reference_id": 1
}
},
"svc-paKzcMJwXghEFJ9UvsdqFU-sid-2": {
"origin": "RECORDING",
"protocol": "SSH",
"recording": {
"session_id": "2",
"verdict": "ACCEPT",
"connection_policy": "ssh2",
"content_reference_id": 2
}
}
}
When the query is the following:
curl --cookie cookies "https://<IP-address-of-SPS>/api/audit/sessions?format=csv&fields=trail_download_link"
The response is the following:
"Key","Audit trail download link"
"svc-paKzcMJwXghEFJ9UvsdqFU-sid-2",""
"svc-paKzcMJwXghEFJ9UvsdqFU-sid-1","https://127.0.0.1/api/audit/sessions/svc-paKzcMJwXghEFJ9UvsdqFU-sid-1/audit_trail"
You can search in the contents of individual connections at the api/audit/sessions/<session-id>/content/?q=<my-search-expression> endpoint.
URL
GET https://<IP-address-of-SPS>/api/audit/sessions/<session-id>/content/?q=<my-search-expression>
Sample request
The following command retrieves those events in the contents of a specific connection that match the search expression(s).
curl --cookie cookies https://<IP-address-of-SPS>/api/audit/sessions/<session-id>/content/?q=<my-search-expression>
NOTE: Make sure that you use the ?q option and that when you use it, you do not leave it empty. Not using the ?q option or an empty ?q will result in an empty "items" list returned in the response.
You can use the Apache Lucene query syntax to create the search expression, but note the following points.
-
You must format the search expression as a URL, and escape special characters accordingly. For example, if your search expression is man iptables, you must escape the whitespace: man%20iptables
-
Do not begin the expression with the * wildcard.
Response
The response contains a list of those events in the contents of the connection that match the search expression(s). The response also contains some meta fields.
If you specified a search expression using the ?q option and the response returns an empty "items" list, that can indicate that:
The following is an example response:
{
"items": [
{
"channel.id": 5,
"end_time": "2017-08-14T10:35:43.957000",
"rank": 2.4756217002868652,
"record_id": {
"begin": 158,
"end": 160,
"for_screenshot": 158
},
"start_time": "2017-08-14T10:35:19.098000",
"trail_id": "12"
}
],
"meta":
{
"href": "/api/audit/sessions/2a620c1cfeb39c537a5e80280283d741/content",
"parent": "/api/audit/sessions/2a620c1cfeb39c537a5e80280283d741",
"remaining_seconds": 599
}
}
items |
|
|
list |
Top-level element, a list containing the details of the matching session. |
|
channel.id |
|
integer |
A reference to the ID of the channel in the session where the event occurred. |
|
end_time |
|
string |
The timestamp of when the content disappeared from the screen. |
|
rank |
|
float |
Indicates the relevance of the match.
If there are several results, the order of them is based on their relevance. |
|
record_id |
|
integer |
The content element's exact position in the audit trail file. |
|
|
begin |
integer |
The identifier of the screenshot in the audit trail file where the content element first appeared. |
|
|
end |
integer |
The identifier of the screenshot in the audit trail file where the content element last appeared. |
|
|
for_screenshot |
integer |
The identifier of the most relevant screenshot in the audit trail file. This is the screenshot on which the event in question is the most clearly visible. For details on how to generate and retrieve the screenshot, see Generate and retrieve screenshot for content search. |
|
start_time |
|
string |
The timestamp of when the content first appeared on the screen and recording started. |
|
trail_id |
|
integer |
The unique identifier of the trail that contains the event. |
In addition, search results can contain the usual meta elements of other endpoints:
meta |
|
JSON object |
Top-level element, a list containing meta information about the response.
For details about the type of information returned, see Message format. |
To generate and download screenshots for a specific content search result, complete the following steps. For details on searching in the content of a session, see Searching in connection content.
-
Perform a content search in a session.
Use a GET request on the endpoint of a specific session, for example:
GET https://<IP-address-of-SPS>/api/audit/sessions/<session-id>/content/?q=<my-search-expression>
For details, see Searching in connection content. If there are search results for the search keywords in the session, the response includes a record_id block, for example:
"record_id": {
"begin": 158,
"end": 160,
"for_screenshot": 158
},
-
Generate a screenshot for the search result.
Note the value of the for_screenshot key in the search response, and use it to generate a screenshot for that particular record_id. POST the value of the for_screenshot key to the https://<IP-address-of-SPS>/api/audit/sessions/<session-id>/_generate?record_ids=<value-of-for_screenshot> endpoint.
-
Download the screenshot.
To download the screenshot in PNG format, GET the value of the for_screenshot key to the https://<IP-address-of-SPS>/api/audit/sessions/<session-id>/screenshots/<value-of-for_screenshot> endpoint.