Hi,
When using Apigee Hybrid (Latest version) for stateful node pool for multi public cloud including the Cassandra in on premise(Hybrid architecture model) & spans across multiple regions and with different public clouds and data centers, what’s the best way to design Cassandra communication across such design?
- Replication factor for every region will 3 (assuming each region has 3 Cassandra), will that not have any performance impacts with such distributed ring design in distributed regions ?
- How do I calculate the number of Cassandra requirements for each regions. Do we have known calculations to derive the number of rings either 3 or 6 or 9 for each region (residing on separate public cloud k8s clusters ) based on ?
- Does compaction is taken care automatically when distributed in such env internally by Apigee Hybrid Cassandra?
- We understand that required port & required firewall to be enabled between all regions for cross communication of stateful nodes across public clouds and on premise regions.
- For stability each region has to undergo Back up and recovery for stateful component for each region ?
- Do we need to consider any other security recommendations, when handling communication between stateful (node rings ?) or only opening port (with required tls per doc) would be suffice from networking perspective Or any additional steps is required ?
- Multiple regions should run the node pool repair say each 10 days @ same time ?.
- Do we have known apigee hybrid 1.8 Cassandra maintenance tasks documented to be considered for such distributed design?
Is there any additional information to be considered in case missed anything ?
Few of docs referred are
Expanding Apigee to multiple regions | Apigee X | Google Cloud
Multi-region deployment on GKE and GKE on-prem | Apigee X | Google Cloud
Configure ports and set up firewalls | Apigee X | Google Cloud
@dino @API-Evangelist @sarthak
LinkedIn
Twitter
