DSP functionality was disabled (scaled down to 0 dspcasa, dspcasa1 and dspintegrator deployments) to prevent an overhead in OI
However this change is affecting alarms anomaly generation for CAPM integration.
How to enable DSP but remove the load on dspintegrator and dspcasa, scoping down the inputs only to CAPM metrics?
DIGITAL OPERATIONAL INTELLIGENCE - 1.3.2
The dspintegrator can get overwhelmed when the volume of metrics becomes larger. In this example we are scoping the processing to only CAPM metrics, this will certainly help the database and processing.
1) Go to Openshift console > Applications > Deployments, locate dspintegrator and open it, go to Environments2) Remove the entries in METRIC_INDEX_ES (you can backup the current value if you want to).3) Set ENABLE_CAPM_METRICS to 'true', by default it is false4) Scale up dspcasa as below:5) Open terminal on dspcasa and view the /opt/dsp/dsp_logs/dsp.log to verify that no errors exist.Additionally, open psql shell connecting to database to see that tables are created in the database. Do this by opening a terminal on postgres pod, launching psql command-line utility, and then listing the databases:psql> \l
backslash 'el'. You should see the dsp database listed. Can connect and verify that schema exists as well now:
psql> \c {dsp_databasename} aopuser
psql> \d+
The listing now should show all tables such as dsp_mpm_green_highway, dsp_mpm_model, dsp_mpm_saved_data, etc.
6) Once above is confirmed, then start 'dspcasa1' pod and wait for it to activate.
7) Finally, 'dspintegrator' pod can be started. After dspintegrator has started up, review the logs tab (or the logs file from Terminal) to verify that content is being processed and no errors are displayed."
- DX OI - Troubleshooting, Common Issues and Best Practices
https://knowledge.broadcom.com/external/article/190815