Projects
A project is the container every session and session template belongs to. Projects keep unrelated work apart inside one HEAT environment. Open Data, then Projects (/admin/projects).
The project list
Summary cards count the total Projects, Total sessions, projects Active in last 7d, and how many are System-managed. A Hide system-managed toggle removes the platform’s own projects from the table so you see only your own. A filter box matches title, description or hex id.
The table lists each project with its session count, last activity and hex id. Select a row to open it.
System-managed projects
The platform creates and owns a small set of projects with reserved titles that start with an underscore. They are marked System and cannot be edited or deleted.
| Title | Purpose |
|---|---|
_statics | Backing sessions for system dashboards |
_logs | Platform log processing |
_diagnostics | Platform diagnostics sessions |
_archived | Sessions moved out of their original project by retention |
_exports | Session export jobs |
Do not create your own project with one of these titles.
The project detail
A project opens on four tabs.
- Overview shows the session count, last activity, storage used and node output count, with the project metadata and the most recent sessions.
- Usage breaks storage and output counts down over time.
- Worker shows the health of the project’s ingestion worker.
- Settings holds the project title, description and the upload buffering configuration.
A Queue on or Queue off control at the top right shows whether the project buffers uploads before storage. With the queue on, many small uploads are coalesced into fewer, larger objects, which keeps storage efficient during high-volume capture. One buffering worker runs per project with the queue on.
Confirming a worker is healthy
When uploads arrive but a session never receives data, check the Worker tab before you suspect the template. It shows whether the project’s buffering worker is deployed and running. A missing or unhealthy worker holds uploads in the queue, which Live Ingestion shows as a session whose output count is not rising.
The default worker memory limit is tight when many large uploads arrive at once. A site that replays large captures should raise the worker memory in the project Settings before a large exercise, not during it.
Related
- Project concept
- Sessions
- Live Ingestion
- Templates