Logstreem
Sandbox

Sandbox

What the documentation sandbox does, what it does not, and how to tell its responses apart.

Simulated, not live

The Try it panels in this documentation call endpoints hosted on this site. They store nothing, make no outbound requests, and are not connected to a Logstreem account.

Telling sandbox responses apart

  • Every response body has a top-level "sandbox": true.
  • Every response carries X-Logstreem-Sandbox: true.
  • Sandbox responses are also X-Robots-Tag: noindex, nofollow.
  • Ids are derived from your input, so the same request always returns the same ids.

The query endpoint is real

One exception is worth knowing about. POST /v1/query in the sandbox actually parses and evaluates your LSQL against a fixed set of 240 events spread over the last hour, across six services and four regions. Change the filter and the rows genuinely change; write invalid syntax and you get the real parse error, with a line number.

Supported in the sandboxNot supported
where with == != > >= < <= =~Parentheses in where
in (…), contains, has, startswith, endswithjoin, make-series, render
extend, project, distinct, countlet bindings
summarize … by with count dcount sum avg min max percentileSub-queries
order by / sort by, take / limitTime-range filters on _time

The sandbox token

Any well-formed token works: it must match lstrm_test_… or lstrm_live_…. The panels default to lstrm_test_sandbox. Send no header, or a malformed one, and you get the real 401 — which is the point.

What is not simulated

BehaviourSandbox
PersistenceNone — ingest accepts and discards
Live tailThe docs show a recorded stream, not a socket
Monitors and notifiersFixtures; nothing is ever evaluated or sent
Rate limitingHeaders are returned, no limit is enforced
Retention and billingNot modelled