Finishing my previous example for multiple IRIS instances, I tried
to compose a local single instance version. The step from the external
Python app to a version using embedded Python seemed to be obvious.
This was a wrong assumption, as some Python libraries just refused installation
into my local Windows-based environment.
After several frustrating and fruitless attempts, I took a step back rethinking the case.
- IRIS offers DeepSee a ready-to-use set of tools for all kinds of charts
- It is ready to present the base data in tables
- Dashboards allow the combination of graphs and tables on a single screen.
- There is no need for an external store.
- And COS allows me any odd trick if ever needed.
The final result

The parts to be served
- My Dashboards are a collection of Table views and Charts
- Tables and Charts are generated from Pivots created in Analyzer
- I use 2 pivots
- based on raw numbers,
- based on $ZLOG(numbers) for better scaling of the chart
- Content for the Pivots comes from 2 separate Cubes
- The Cubes are designed with DeepSee Architect from a persistent class
- The class is refreshed for each view by a Class-Query in %SYS
- To avoid useless Journal records, I use a temporary table in IRISTEMP
- Finally, a Portlet provides the actual timestamp of the snapshot
The processing steps
- A CSP page is called by its URL
- In method OnPreHTTP
- The System Query fills the temp table from %SYS
- Then both cubes are rebuilt
- Finally, the Dashboard is launched by server-side redirection
- The content of the initial CSP page is irrelevant and therefore empty