Skip to main content

Splunk

OrbisID can send its findings, privileged accounts and KRI values to Splunk through the HTTP Event Collector (HEC), so a security analyst sees OrbisID context (not in a vault, no owner, high risk) inside Splunk searches and alerts. This is available in the Pro and Enterprise editions, and the Administrator role sets it up under Administration > Integrations > SIEM Export. The same page configures the Microsoft Sentinel export.

Changes are sent in the minutes after a scan or a KRI snapshot, in batches, and a delivery that fails is retried by the next run so nothing is skipped.

What is sent​

Three kinds of event, each with its own sourcetype (the prefix is configurable and defaults to orbisid). An event is sent when the item changes, so the index holds the history; use dedup to read the latest state of each item.

SourcetypeSent whenOne event per
orbisid:findingA finding is created or changes (status, owner, severity, due date)Change to a finding
orbisid:privileged_accountA privileged account is detected or updated by a scan or an overrideUpdate to a privileged account
orbisid:kriA KRI snapshot is taken (up to 10 unsent snapshots per run, oldest first; the first run sends only the latest)KRI per snapshot

OrbisID never sends a password, key or secret. Findings and accounts carry names and the owner's name and email, so restrict the index to people who may see that.

Each event is a JSON object. The fields are the same as in the Microsoft Sentinel export:

  • Findings: EventTime, FindingId, KriCode, KriName, EntityType, EntityId, EntityName, SystemName, Severity, FindingStatus, OwnerName, OwnerEmail, OpenedAt, FirstSeenAt, DueAt, ResolvedAt, ReopenCount.
  • Privileged accounts: EventTime, AccountId, AccountName, SystemName, SystemOsType, AccountType, NhiSubtype, Enabled, OwnerName, OwnerEmail, OwnerUsername, Ownership (OWNED, ORPHANED or NO_OWNER), PamRiskLevel, InPamVault, LastUsedAt, LastUsedSource, LastDetectedAt.
  • KRI values: EventTime, SnapshotId, KriCode, KriName, GrossValue, AdjustedValue, DenominatorValue.

The Splunk event time is EventTime. A value OrbisID does not know is left out. The event also carries host (default orbisid), source orbisid and, if you set one, the index.

Set it up in Splunk​

  1. Create the index that will hold the data (for example orbisid). In Splunk Cloud Platform an index must exist before a token can use it.
  2. Turn on HEC (Splunk Enterprise: Settings → Data inputs → HTTP Event Collector → Global Settings) and decide on TLS: Enable SSL makes HEC listen over HTTPS. The default port is 8088.
  3. Create a HEC token (New Token). Under Select Allowed Indexes choose the index. Copy the token value.
  4. Note the HEC URL. For Splunk Enterprise it is https://<your-host>:8088. For Splunk Cloud Platform it is https://http-inputs-<your-stack>.splunkcloud.com (port 443, or 8088 on a free trial); on Google Cloud and AWS GovCloud the host prefix differs, so use the address Splunk shows for your stack.

OrbisID requires an https HEC URL because the token travels in every request.

Set it up in OrbisID​

Open Administration > Integrations > SIEM Export → Splunk and enter the HEC URL (with or without /services/collector/event), the token, and optionally the index, the host value and the sourcetype prefix.

  • Verify the TLS certificate is on by default. Splunk's own HEC certificate is self-signed out of the box, so OrbisID will refuse it until Splunk uses a certificate that OrbisID trusts. You can turn verification off; OrbisID then accepts any certificate and cannot tell it is talking to your Splunk, so use that only on a trusted network.
  • Send test event sends one real event with the sourcetype <prefix>:test to the index. It checks the URL, the token and the index together.
  • Send now sends what changed straight away. The first run sends all findings and privileged accounts and the latest KRI snapshot only; later runs send only changes.
  • The panel shows the last run, the last success and the last error. After a failure the same events are sent again by the next run.

The token is stored encrypted and is never shown again; leave it blank to keep the saved one.

If delivery fails​

The error shows Splunk's HEC code. The common ones are:

HEC codeHTTPMeaning
1, 4, 21, 22403 or 400The token is disabled or wrong
2, 3401The authorization was not accepted
5, 6, 12, 13400Splunk could not read the event data
7400Incorrect index: it does not exist or the token is not allowed to write to it
9, 18 to 20, 23, 26, 27503 or 429Splunk is busy or shutting down; OrbisID waits and tries again, up to four times

Search it​

These searches are starting points; adjust the index and sourcetypes to your setup. If the fields do not appear in search, set the three sourcetypes to extract JSON (KV_MODE = json in props.conf).

Open findings by severity:

index=orbisid sourcetype="orbisid:finding"
| dedup FindingId
| search FindingStatus IN ("OPEN", "IN_PROGRESS")
| stats count by Severity

Privileged accounts that are not in a vault or have no active owner:

index=orbisid sourcetype="orbisid:privileged_account"
| dedup AccountId
| search Enabled=true (InPamVault=false OR Ownership!="OWNED")
| table AccountName SystemName InPamVault Ownership OwnerName PamRiskLevel

Enrich Windows logons with OrbisID context. First keep the latest privileged accounts as a lookup, with a scheduled search such as:

index=orbisid sourcetype="orbisid:privileged_account"
| dedup AccountId
| eval user=lower(AccountName)
| table user InPamVault Ownership OwnerName PamRiskLevel
| outputlookup orbisid_privileged_accounts.csv

then add the context to the events you already collect:

index=wineventlog EventCode=4624
| eval user=lower(Account_Name)
| lookup orbisid_privileged_accounts.csv user OUTPUT InPamVault Ownership OwnerName PamRiskLevel
| search PamRiskLevel=*
| table _time user host InPamVault Ownership OwnerName PamRiskLevel

The account name field differs between logon sources and add-ons; use the one your events carry.

Non-human identity hygiene over time:

index=orbisid sourcetype="orbisid:kri" KriCode IN ("ORPHANED_NHI", "NHI_WITHOUT_OWNER", "EXPIRED_ACTIVE_CREDENTIALS", "LONG_LIVED_CREDENTIALS")
| timechart span=1d max(AdjustedValue) by KriName

A simple dashboard​

Create a new classic dashboard, switch to Source and paste:

<dashboard version="1.1">
<label>OrbisID posture</label>
<row>
<panel>
<title>Open findings by severity</title>
<chart>
<search>
<query>index=orbisid sourcetype="orbisid:finding" | dedup FindingId | search FindingStatus IN ("OPEN","IN_PROGRESS") | stats count by Severity</query>
<earliest>-90d@d</earliest>
<latest>now</latest>
</search>
<option name="charting.chart">pie</option>
</chart>
</panel>
<panel>
<title>Unvaulted or ownerless privileged accounts</title>
<table>
<search>
<query>index=orbisid sourcetype="orbisid:privileged_account" | dedup AccountId | search Enabled=true (InPamVault=false OR Ownership!="OWNED") | table AccountName SystemName InPamVault Ownership PamRiskLevel</query>
<earliest>-90d@d</earliest>
<latest>now</latest>
</search>
</table>
</panel>
</row>
<row>
<panel>
<title>Non-human identity KRIs</title>
<chart>
<search>
<query>index=orbisid sourcetype="orbisid:kri" KriCode IN ("ORPHANED_NHI","NHI_WITHOUT_OWNER","EXPIRED_ACTIVE_CREDENTIALS","LONG_LIVED_CREDENTIALS") | timechart span=1d max(AdjustedValue) by KriName</query>
<earliest>-90d@d</earliest>
<latest>now</latest>
</search>
<option name="charting.chart">line</option>
</chart>
</panel>
</row>
</dashboard>

Threat Detections over syslog​

Threat Detections from Endpoint Sensors can already be forwarded as CEF syslog messages (see Threat Detections); that is separate from this export, and Splunk can receive both.

Splunk documentation​

OrbisID's use of the collector follows Splunk's Format events for HTTP Event Collector (endpoint /services/collector/event, Authorization: Splunk <token> header, event envelope, time in <sec>.<ms>, concatenated JSON events in one request), Troubleshoot HTTP Event Collector (the HEC status codes above) and Set up and use HTTP Event Collector in Splunk Web (port, Splunk Cloud address, allowed indexes).