Currently, we are using the PipeCD agent to deploy a large number of Terraform applications. However, the agent starts running all deployments that have been triggered without considering its processing ability. As a result, our PipeCD agent often gets stuck when multiple deployments are triggered simultaneously.
Therefore, I wish we had a way to specify a limit number in the PipeCD agent configuration to prevent it from burning out.
Some other considerations to take into account:
The limit number should be applied to each type of application (e.g., Terraform apps may require more resources to handle than other types of apps).
Deployments triggered by plan-preview should also be counted in the limit.
What would you like to be added:
Why is this needed:
Currently, we are using the PipeCD agent to deploy a large number of Terraform applications. However, the agent starts running all deployments that have been triggered without considering its processing ability. As a result, our PipeCD agent often gets stuck when multiple deployments are triggered simultaneously.
Therefore, I wish we had a way to specify a limit number in the PipeCD agent configuration to prevent it from burning out.
Some other considerations to take into account:
The limit number should be applied to each type of application (e.g., Terraform apps may require more resources to handle than other types of apps). Deployments triggered by plan-preview should also be counted in the limit.