Security Incident Response cloning best practices

mariedutton
Tera Contributor

Hello SIR Experts,

Does anyone have a tried and true/battle-tested documented procedure for what to exclude/preserve/cleanup in a clone from production down to the lowers to prevent security incident data from being exposed anywhere other than production. We want to be able to let sys admins have sn_si.admin in the lowers for development purposes, so the production data needs to be excluded. Claude and I created a procedure with what I think are the logical tables in the sn_si and sn_ti and other sn_sec_cmn scopes and a script to clean up  attachments, journals and integration creds. It feels pretty complete to me. But I figured it wouldn't hurt to ask if anyone out there had any experience with this, gotchas to pass on, or even a "no warranty-as-is" process doc that has been used IRL.

Thank you,

Marie

1 REPLY 1

kmohammed
Tera Guru

Your approach is on the right track, and the most important refinement is this: treat exclusion as the first control, not the only control. The procedure is only safe if no one with lower-environment access can reach production SIR data before post-clone verification is complete.

 

A workable documented procedure usually has these four parts:

 

1. Exclude transactional SIR data at clone time.

Use clone data exclusions for the transactional tables in the sn_si, sn_ti, and sn_sec_ccm scopes: security incidents, their tasks, observables, indicators, and related transactional records. Review the list before every clone, because new tables and extensions added by an update can quietly fall outside an old exclusion list. Do not rely on a list that has not been reviewed since the last release.

 

2. Preserve configuration, not data.

Keep the definitions that developers actually need: playbook definitions, SIR configuration, assignment and routing configuration, and other setup records that contain no incident content. That distinction is what makes the lower environment useful for development without carrying production incidents with it.

 

3. Clean up what exclusion alone does not catch.

This is the gotcha most procedures miss. Incident content can survive in tables outside the SIR scopes, so your post-clone script should also cover attachments, work notes and journal entries, and email records that reference SIR records, plus integration credentials. A table exclusion that leaves an attachment behind has not actually excluded the incident.

 

4. Neutralize the target before anyone logs in.

Disable outbound integrations and scheduled jobs that could push data back to production systems or to third-party tools, and rotate or reset credentials in the target. A cloned integration pointing at a production tool is a data path, not just a configuration detail.

 

Then add a hard gate before access is granted:

 

- Do not grant sn_si.admin, or any broad lower-environment access, until verification is done.

- Verify with counts and spot checks: confirm the SIR transactional tables are empty or contain only test data, search for a known production incident number, and check attachments and journal entries for SIR references.

- Only after that check passes should developers be given access, and development should use synthetic or masked test incidents rather than restored production records.

 

If you document it that way, with the exclusion list, the cleanup script, the neutralization steps, and the verification gate all as required steps rather than optional ones, you have a procedure that holds up even when someone adds a new SIR table or a new integration later. The risk is rarely the one table everyone remembers; it is the attachment, note, or integration that was not in the original list.