Stratusphere API
The Stratusphere Hub serves a read-only data API for what Stratusphere collects. Whether it can change Hub settings is unverified. Every Hub serves its own full option reference, examples, and per-inspector column catalog at https://<hub>/lwl/api—a real Hub's built-in documentation always wins over anything recorded here.
Request shape and authentication
From Liquidware's Using the API and Accessing the API guides:
- Endpoint:
https://<hub>/lwl/api?json={...}, called withGETorPOST. - Required fields:
inspector(the data table ID) andbasis(the group-by dimension, such asusersreturninguser_idanduser_name, ormachinesreturningnode_idandnode_name). Without a time window, the default period isyesterday. - Authentication: Access is gated by a client IP allow list (Hub Administration > Configuration > Security Settings), optionally paired with HTTP Basic authentication as
username@domain. The built-inssadminaccount cannot call the API; create a dedicated read-only API user (hub-setup.md). - Output formats: JSON, CSV, or Excel.
Inspectors and parameters seen in public samples
From Liquidware's public Stratusphere-UX SE PowerTools samples (reference only, no license; see community-projects.md) and SE Field Articles:
| Inspector ID | Dataset | Typical basis |
Notes |
|---|---|---|---|
0 |
Users / summary | users, machines |
Overall user or machine summary metrics |
1 |
Applications | applications, users, machines |
Application load counts, load time, CPU/memory by process |
2 |
Connections / remoting | users, machines |
Session protocol and connection metrics |
3 |
Virtualization hosts | hosts |
Hypervisor / ESX host metrics |
4 |
Machine inventory | machines |
OS version, hardware make/model/type, CPU cores/speed, disk size, allocated memory, last boot |
10 |
Login details | users, machines |
Logon breakdown and timing |
Common request parameters:
- Time window:
date(today,yesterday,2 hours,24 hours,1 day),custom_start_date/custom_end_date(YYYY-MM-DD),start_time/end_time(hour integers, used for business-hour slicing),resolution(Day-Hour). - Projection and filtering:
columns(array or comma-separated list of column names),filter(expressions such asapp_load_count gt 0,user_name not like '...',d_id eq 5308 OR ...),only_process(user),node_group_namewithunbasis. - Paging and formatting:
limit(0for all rows),sort_col,output(1withoutput_namefor background file generation),header,csv_eof.
Ecosystem and reporting patterns
- Background generation for large queries: Large API responses are built asynchronously on the Hub. Liquidware's SE automation article retries after 30 seconds with a 180-second timeout.
- Power BI integration: Liquidware's 2019 blog post points Power BI's Web connector directly at the Hub API URL, which requires an on-premises data gateway, a trusted Hub certificate, and the Hub remaining online for every refresh. Exporting to a normalized local dataset first (the pattern used by
ux-assess, see spotcheck-and-sizing.md) avoids gateway and certificate dependencies and feeds Power BI the same pseudonymized dataset as written reports. - Built-in Executive Summary report: Stratusphere 6.7+ includes an Executive Summary UX custom report comparing KPIs across two time periods inside the Hub.
Gotchas
- Self-signed TLS certificate: The Hub uses a self-signed certificate by default. Trust that certificate on the exporting machine instead of disabling TLS verification globally (Liquidware's older
BatchReport.ps1sample disables TLS checks). - Inspector Excel export row cap: Exporting an Inspector grid directly to Excel from the UI caps at 99,999 rows per sheet (unverified). Use the API with
limit: 0or slice by day/hour for larger estates. - Unverified column names and units: Until confirmed against a live Hub's
/lwl/apipage, exact column names for login duration and composite UX score, the units ofmemory_allocatedandapplication_load_time_avg, and whetherresolution: Day-Houremits a clean timestamp column remain unverified. - PII in exports: Raw API responses contain real user and machine names. Keep raw exports in git-ignored
output/directories and pseudonymize with HMAC before data leaves the customer environment.