Policy 11/05/22-25/05/22, Policy 25/05/22-08/06/22, Policy 08/06/22-22/06/22, Policy 2022-06-30->2022-07-13
As a ONAP developer, I would like to perform a spike to determine the work needed in drools-pdp for redundancy support. Basically that means increasing the replicaSet from 1 to at least 2 and verify all the operations are working as expected. Following are the main areas to look at:
- Sending registration & heartbeat requests to PAP.
- Deployment/Undeployment of policies.
- Reading events from DMaaP topic in parallel.
- Processing events in parallel.
- Reading response of an execution request sent to actors.
- Updating operations in policy database.
- Persisting logs from multiple PDP instances in a single k8s PV.
- Fetching prometheus metrics from multiple PDP instances.
- Fetching Statistics for multiple PDP instances
- Distributing the load between the multiple PDP instances.
- What is needed to make the PDP stateless, thereby converting statefulset to deployment in oom charts.
- Check the possibility of performing rolling update of the k8s pods.