Runtimes Based Inventory (RBI) is a service that manages data for Runtimes workloads. It is intended to serve the same role for Runtimes that Host-Based-Inventory does for RHEL Insights.
Both components come from the same Quarkus codebase, with just a different entrypoint into the container used.
This will be split into 2 sets of pods (in Staging & Prod) to allow independent scaling of the two operations.
Each set will run a slightly different image.
This component provides a simple REST interface to read data from the customer's dataset.
A Kafka listener that writes data into an RDS postgres instance.
Routes into RBI are controlled by 3Scale.
console.redhat.com/api/runtimes-inventory/v1/openapi.json Has a description of the API (autogenerated)
console.redhat.com/api/runtimes-inventory/v1/workloads Will expose an access-controlled list of UUIDs that identify individual Runtimes workloads instances, and the RHEL host they are running on. Will support date parameters (defaulting to last 24 hours) in a future version.
console.redhat.com/api/runtimes-inventory/v1/workloads/ The details for a specific workload.
Access control for both routes is “customer can only access their own data, that matches their accountId & orgId”
RBI uses a Postgres database hosted on AWS. If the database is unavailable, this will be a major outage for the application. As soon as the database is available again, the application will recover by itself.
Migration and upgrades are handled Flyway scripts. Quarkus uses Flyway as a standard DB migration / versioning tool - https://flywaydb.org/
It’s widely deployed at Red Hat & is used on Insights by e.g. notifications-backend.
RBI uses the main consoledot Kafka and dedicated topics. RBI is agnostic to the type of Kafka used.
Store and process information related to the usage of EAP and other Runtimes products - to allow us to understand customer usage.
The database stores all application-related data. If unavailable, the application will have a significant outage. Past workload statuses will be lost in case of data loss.
Tbd
| Component | CPU (request/limit) | Memory (request/limit) | Storage |
|---|---|---|---|
| rbi-rest | 2x 300m/1 | 2x 256Mi/256Mi | - |
| rbi-topic-read | 2x 300m/1 | 2x 256Mi/256Mi | - |
| RDS | db.m6g.large | db.m6g.large | 50Gi |
We don't expect significant changes in the capacity requirements in the next 6 months.