nano SIEM
Search Commands

tree

Build a parent–child tree from events: process trees, web referrer chains, or any parent/child field pair.

Use tree to inspect parent–child relationships in your matching events. You can build a collapsible process tree, web referrer chain, or tree from your own field pair.

Syntax

Presets:

... | tree process [root=<pattern>]
... | tree web [root=<pattern>]

One field with a parent field:

... | tree <field> parent=<parent field> [root=<pattern>]

Full form:

... | tree parent=<field> child=<field> label=<field> [detail=<field>] [prevalence=<field>] [root=<pattern>]
ArgumentDescription
parentField holding the parent's identifier. Required except in presets.
childField holding the node's own identifier.
labelField shown as the node name.
detailOptional field shown under each node, such as a command line.
prevalenceOptional field the tree view uses for rarity badges.
rootOptional pattern. Shows only subtrees whose root node's label matches.

In the one-field form, the field is both the identifier and the label. by, depth=, maxchildren= and format= are accepted and ignored.

Presets

Presetparentchildlabeldetailprevalence
tree processactor.process.uidprocess.uidprocess.nameprocess.cmd_lineprocess.file.hashes.value
tree webhttp_request.referrerhttp_request.url.url_stringhttp_request.url.hostname—http_request.url.hostname

tree process fits Process Activity events (class_uid=1007), where the parent is the acting process. tree web fits HTTP Activity events (class_uid=4002).

root patterns

PatternMatches labels that…
"cmd.exe"equal cmd.exe (case-insensitive)
"*shell*", "power*", "*.exe"contain, start with or end with the text
/wscript|cscript/match the regular expression (case-insensitive)

The full tree is built first, then filtered to matching subtrees with all their descendants. The ancestors of a matching node are shown above it.

How it works

  • Nodes whose parent isn't in the result become roots.
  • The tree is built from at most 10,000 parent/child events. When more events match, the tree is partial and the result says so: Tree built from N parent/child events; more matched, so the tree is partial. Narrow the search or time range for the full tree.
  • In the tree view, click a node label to add it as a filter, or right-click to exclude it.
  • tree can't be used in dashboard panels.

tree also works after a lookup, prevalence, inputlookup, resolve_identity, or ai stage.

Examples

Process tree for one host:

class_uid=1007 device.hostname="ws-it-083.corp.local" | tree process

Only the subtrees under PowerShell or cmd.exe on that host:

class_uid=1007 device.hostname="ws-it-083.corp.local" | tree process root=/powershell|cmd/

Everything Office applications spawned, across hosts:

class_uid=1007 | tree process root=/winword|excel|outlook|powerpnt/

Web referrer chains for one client:

class_uid=4002 src_endpoint.ip="10.1.3.87" | tree web

Referrer chains that start at one site:

metadata.log_source=conduit_proxy | tree web root="*google.com*"

Process-name tree (which process names start which), without per-process identifiers:

class_uid=1007 | tree process.name parent=actor.process.name

Full form with explicit fields:

class_uid=1007 | tree parent=actor.process.uid child=process.uid label=process.name detail=process.cmd_line prevalence=process.file.hashes.value

What's next

  • sequence: ordered patterns
  • lateral: movement between hosts
  • prevalence: rarity of hashes, domains and other artifacts
  • asset: investigation view for one host

On this page