Alternative architectures
Itential recognizes the need to be flexible. The information detailed on this page includes suitable alternatives to the validated designs. While these exceptions are not part of the architecture testing cycle, they provide some advantages and are considered acceptable architectures.
Colocate Redis and MongoDB
In non-production environments, Redis and MongoDB can be colocated on the same VM to reduce the size of the infrastructure footprint. Although this reduces the infrastructure footprint, it is not well understood what strains this architecture will place on the components, nor what kind of workloads could trigger issues. This architecture should be used only for non-production environments.
All-in-one architecture
In development environments, Itential Platform, Redis, and MongoDB can be installed on the same VM. Although this reduces the infrastructure footprint, it is not well understood what strains this architecture will place on the components, nor what kind of workloads could trigger issues. This architecture should be used only for development environments.
All-in-one does not include IAG on the same server. Such architecture is not supported as Itential Platform and IAG applications use conflicting versions of Python.
Server specifications
Disk references to pronghorn (seen in older deployments) should be changed to itential.
All-in-one
IAG server
MongoDB Atlas
Some customers may not want to run and maintain their own MongoDB cluster. A suitable alternative to running a MongoDB cluster is to use MongoDB Atlas. Running Atlas does not remove the care and attention that is required when planning this architecture. Itential Platform is very sensitive to network latency. A recommended practice is to keep your database close to your instance of Itential Platform, as this will help prevent unwanted latency. If running everything in AWS, then you must ensure that all of the components reside in the same Virtual Private Cloud (VPC).
Elasticache
Another suitable alternative for customers who may not want to run and maintain their own Redis cluster is to use AWS Elasticache. Running Elasticache does not remove the care and attention that is required when planning this architecture. Itential Platform is very sensitive to network latency. A recommended practice is to keep Redis close to your instance of Itential Platform to help prevent unwanted latency. If running everything in AWS, then you must ensure that all of the components reside in the same Virtual Private Cloud (VPC). Elasticache must be configured to use Redis. It is not possible to use a distributed memory object caching system (memcache).
