... | @@ -101,3 +101,5 @@ env = K8sParallelEnvV0(config_file, agent_deploy_files, kwargs) |
... | @@ -101,3 +101,5 @@ env = K8sParallelEnvV0(config_file, agent_deploy_files, kwargs) |
|
* **config_file**: path to global yaml config file containing the environment's parameters
|
|
* **config_file**: path to global yaml config file containing the environment's parameters
|
|
* **agent_deploy_files**: a list of paths to agent yaml deployment files containing Pods, Services, Deployments etc.
|
|
* **agent_deploy_files**: a list of paths to agent yaml deployment files containing Pods, Services, Deployments etc.
|
|
* **kwargs**: keyword arguments that can be used to override parameters found in the config file. If a keyword argument is not found in the config file, it is skipped.
|
|
* **kwargs**: keyword arguments that can be used to override parameters found in the config file. If a keyword argument is not found in the config file, it is skipped.
|
|
|
|
# Graphs application variant
|
|
|
|
A variant of this environment was created to create an application that performs graph operations. The service graph consists of three microservices. The first one receives a number as the input size and creates two graphs with the given size. Afterwards, it creates a composition of these two graphs. The second microservice receives the unified graph and performs link prediction adding the calculated edges if their resource allocation index is above a certain threshold. The third microservice receives the previous graph and calculates the degree centralities for all nodes. Since the requests are all performed asynchronously, two auxiliary microservices were deployed, one at the entry and one at the exit point of the application graph. The first one marks the beginning of the user's request and the last one calculates the total time taken to serve the request and exposes it to Prometheus. |
|
|
|
\ No newline at end of file |