-
Start monitoring a local network with Flow processing and verify usage of the UI/Grafana servers, ensuring everything works.
-
Use relative/percentage size for the Java Heap, based on Pod resources (instead of fixed values) for OpenNMS.
- For instance,
-XX:MaxRAMPercentage={{ .Values.opennms.jvm.heapPercentage }}
- We could let the user choose between discrete assignments via Xms/Xmx or percentage.
- For instance,
-
Evaluate the idea of having custom entry point scripts replacing the initialization scripts.
- The less invasive option to expand our possibility without building custom images.
- There are limitations with
confd
in OpenNMS, besides other restrictions inside theentrypoint.sh
script in OpenNMS and Sentinel that prevents enabling certain features.
-
Improve Helm Chart for OpenNMS and relatives (no external dependencies).
- Make readiness/liveness probes configurable (DB migration time could impact behavior). Or keep them fixed and expose settings for the startup probe.
- Use the
lookup
function to ensure that theStorageClass
exists and fail if it doesn't. Or use it to only create it if it doesn't exist (and reduce requirements).
-
Learn how the GKE Ingress works, in case the Nginx Ingress is not an option.
-
Consider
NetworkPolicies
to isolate resources on a given namespace. -
Test the solution against a secured Grafana Loki server.
-
Build Terraform recipes for the Cloud Infrastructure resources.
- Private container registry
- Kubernetes cluster
- SQL Service for PostgreSQL