Running Federated Search from Fusion SOAR

Run FSQL against every connected source from a Fusion workflow

📘

TL;DR

The app exposes federated search as a Fusion SOAR action. It is read-only, returns rows plus flags telling you whether the result was complete, and must finish within 30 seconds.

Adding the Action

  1. Next-Gen SIEM → Fusion SOAR → Workflows, create or edit a workflow.
  2. Add the federated search action provided by the Query app.
  3. Supply the FSQL statement, and optionally a maximum row count.

Reference trigger fields with the workflow's data pills, so a detection's hostname or hash flows into the statement.

🚧

Use the pills rather than typing references

A mistyped data reference does not error — it is stored as a literal string and the workflow runs with it.

What You Can Run

Non-mutating statements only, such as: QUERY, SUMMARIZE, ANALYZE, ...

QUERY authentication.** WITH %username = 'jdoe' SINCE 6h LIMIT 20

What It Returns

For a QUERY, the matched events plus:

OutputMeaning
Row countEvents returned
SourcesDistinct platforms that contributed
Unattributed rowsRows whose source could not be determined from the event
TruncatedWhether the result was capped
Total rows availableHow many existed when it was capped
PartialWhether some connectors failed while others succeeded
Search and trace idsFor correlating with Query platform logs

EXPLAIN and other non-QUERY commands return their payload directly rather than as rows.

📘

Branch on partial if the workflow makes decisions from the result. A search can succeed and still be incomplete.

Limits

30 seconds per call, which is a Fusion SOAR limit. To stay within the limits, you can bound the time to a narrow window, pin an entity, and use a LIMIT (the search returns as soon as the limit is satisfied from any source, so it bounds latency as well as payload). Note that omitting a time clause searches the last 24 hours.

LIMIT only shortens a search where matches are plentiful; a selective predicate over a sparse window can still run long. Narrow the time window instead.

Row cap. Bounds what the action returns, not what the query executes — a LIMIT clause is still the way to bound the search itself. When the cap trims a result, the action reports how many rows were available.

Response size. Falcon limits response sizes to toughly 1 MB. A broad OCSF event could be large; hence the action trims to fit (and says so).

Example Patterns

  • On case creation, search recent authentication activity for the named account and attach the result.
  • After a detection fires, search connected threat-intelligence sources for the hash and branch on the outcome.
  • Search network_activity for the host named on a detection over the preceding hours.
📘

For several lines of inquiry and a written conclusion, use an agentic investigation instead — see Auto-Investigating Detections.

What's Next


Did this page help you?