This page provides an overview of the hardware requirements for Dynatrace Configuration as Code via Monaco.
The main hardware requirement of the Dynatrace Monaco CLI is memory.
When executing in a container with a hard memory limit or limited physical hardware, make sure to set a reasonably high memory limit or ensure that swap space is available.
The required memory is mostly defined by the number of configurations and the size of their JSON template files. The following tables give you an idea of memory requirements for sample project sizes; your project might need more or less memory.
Note that the CLI in version 2.9.0 and earlier has higher memory requirements.
Dynatrace Monaco CLI version 2.9.1+
For larger projects, follow these rules for estimating the required memory for a deployment:
Both of these measures are better applicable for larger projects, where memory needs unrelated to configuration are less noticable.
You can configure a soft memory limit for the Dynatrace Monaco CLI by setting the GOMEMLIMIT environment variable.
GOMEMLIMIT is a numeric value with an optional unit suffix of B, KiB, MiB, GiB, or TiB.
If no suffix is supplied, the value is assumed to be in bytes (B).
For example, to set a limit of 3 gibibytes, set the environment variable as GOMEMLIMIT=3GiB.
GOMEMLIMIT=3GiB monaco deploy manifest.yaml
As a soft limit, this limit can be exceeded if your deployment needs more memory. Setting too low a limit, however, will result in increased runtime, as more time is spent to free memory when the limit is exceeded. See the table above for estimated memory requirements.
Dynatrace Monaco CLI version 2.8.0+ A default memory limit of 2GiB is applied. To change it, set GOMEMLIMIT.
Dynatrace Monaco CLI version 2.7.0 or earlier No default memory limit is applied. The Dynatrace Monaco CLI may consume excessive memory. To apply a memory limit, set GOMEMLIMIT.
Available CPU mostly impacts deployment time needed and, unlike memory, doesn't impose hard limits; on more limited hardware, deployments simply take more time.
For example, deploying from a container limited to a single vCPU takes significantly longer than one with several available vCPUs or running directly on a system with a multi-core processor.
The following tables give a rough overview of how the number of configurations impacts deployment times. This merely illustrates the relationship between project size and deployment time and probably won't match deployment time on your specific hardware.
Dynatrace Monaco CLI version 2.7.0+
This table demonstrates that deployment time doesn't grow linearly with the size of your projects. This is because of the behavior of the Dynatrace API as the number of requests and existing configurations increases—this causes increased rate limiting and processing time when adding configurations.