.svg?alt=media&token=0f6df7c7-5b98-4a62-98b8-a54e38a8329e){kind=logo}
FAQs
Popular and frequently asked questions about DuploCloud
DuploCloud supports the following public cloud providers:
- Amazon AWS
- Microsoft Azure
- Google Cloud
- On-Premises
Use these FAQ documents to quickly find answers to popular questions about using AWS, Azure, and GCP with DuploCloud.
No. DuploCloud is a self-hosted solution deployed within the customer's cloud account. This hosted solution provides the customer with a SaaS-like experience. If the customer desires, DuploCloud can provide a fully managed service to maintain uptime, provide updates, and supply ongoing support.
DuploCloud is a self-hosted single-tenant solution deployed within the customer's cloud account. The software runs in a virtual machine (VM) and the VM derives permissions to call the cloud provider using the VM's permissions. For example, in AWS, permissions are derived via an instance profile. In Azure, permissions are derived via managed identity.
The DuploCloud VM and DuploCloud Portal are secured, as is any other workload in the cloud. In addition to SSO login for portal access, the VM runs behind a VPN. Therefore, only internal users can load the portal when connected to a VPN.
DuploCloud is running in your own cloud account, along with your workloads. DuploCloud is a provisioning system, so stopping DuploCloud does not impact any of your applications and cloud services.
The following is a list of automation constructs managed by DuploCloud, and a summary of what you need to do to maintain them directly, instead of through DuploCloud.
- 1.Cloud Provider Configuration (Terraform): This involves various cloud services, IAM roles, Security groups, VPC, etc. DuploCloud can export your latest cloud configuration into native Terraform code and state files. Once exported, you maintain the configuration.
- 2.Kubernetes: All applications and configurations that have been deployed in K8s are available in the form of deployments, StatefulSets, DaemonSet, K8s Secrets, ConfigMaps, etc. One can run
kubectlcommands to export configurations as YAML files and continue to maintain them in the future. - 3.Compliance monitoring: DuploCloud uses a third-party SIEM solution called Wazuh. Wazuh is an open-source software platform running in an independent VM in your cloud account, and you have full permission to retain it "as-is." However, going forward, any new systems that need compliance monitoring need to be integrated into the SIEM by you.
- 4.Diagnostics tools: Tools include Prometheus, Grafana, and Elasticsearch. All of these tools are open source and run in your cloud account. You can continue to manage them directly.
If DuploCloud is down, it's similar to having an unavailable DevOps engineer. If you need to opt out of DuploCloud, you are replacing your DevOps management. DuploCloud is neither a PaaS nor a hosted solution, which hosts your workloads.
Absolutely! In fact, more than half of our customers have no DevOps team. With our managed service offering, we handle your deployments, act as the first line of defense for any issues, and are constantly involved in daily tasks like CI/CD updates, etc.
The DuploCloud team acts as your extended DevOps team, assisting with white glove environment setup and daily operations with 24x7 Slack and/or email support. We cover what is supported in the DuploCloud platform, as well as assist with your cloud provider's requirements.
After the initial onboarding of the platform, we recommend that you engage the DuploCloud team as your second line of defense by setting up an internal triage process of your own. We can assist you in setting up this process.
DuploCloud is your extended DevOps team! We are available 24x7 on your Slack channel, by phone, and by email.
Yes. Many times this is required because DuploCloud may not support all cloud features or configurations. Direct changes to your cloud account can be categorized within the following groups:
DuploCloud labels the resources and configurations it manages. If independent changes are being made directly in your cloud provider, DuploCloud does not interfere. However, DuploCloud still monitors changes for compliance and alerts you to any non-compliant configurations.
For resources that DuploCloud does not directly manage, you make the change directly in your cloud provider. For example, if you add additional forwarding rules to a load balancer created through DuploCloud, DuploCloud does not interfere with the new configuration that you created.
For resources that DuploCloud manages, DuploCloud automatically detects the conflicts and either reverts the changes or raises an alert about the inconsistency.
DuploCloud provides flexibility when a feature that is not supported by DuploCloud can be programmed directly in your cloud provider. If you are using Terraform, then the DuploCloud provider, as well as the native cloud provider, can be used in tandem. For an example of this use case, see https://duplocloud.com/white-papers/devops/#PAAS.
Yes. DuploCloud's Web UI is a no-code interface for DevOps. You do not need to know IaC or have any cloud expertise to operate it. You simply need to understand the basic constructs in DuploCloud by reading the product documentation.
The answer depends on the following key factors:
Especially in a rapid-growth company environment, where architecture is constantly evolving and services are being constantly updated, there may be a desire to let developers self-service and move fast. This would be a case for no-code. On the other hand, in cases where there is an established operations organization, with centralized requirements, then a low-code Terraform solution may be a better option.
Note that Terraform is a client-side scripting tool and a single-user system, so the scope of a project is limited to one person operating at one time. This can incur constraints if a project has many components and two people cannot operate at the same time, even if they are dealing with completely independent constructs.
Terraform projects typically have a broad scope with multiple components. At times, you need to make small targeted changes (for example, a health check URL change). But when the change is being executed, there may be other drifts and the user is forced to resolve them. This can be inconvenient and often the user will make the change in the UI, resulting in further configuration drifts.
About half of our customer base uses no-code, while the other half uses Terraform. Ironically, software developer-centric companies prefer no-code because it enables engineers to be agile and focus on their application code. By contrast, in DevOps-centric organizations, the low-code Terraform provider is often used.
DuploCloud licenses usage is calculated based on the services managed by DuploCloud. The service usage is counted in terms of units, with a unit defined as below:
- A host is counted as 1 unit. (example: EC2 instance, Azure VM)
- A serverless function or service is counted as 1/4 unit (example: Lambda function)
- A serverless application is counted as 1/2 unit (example: AWS ECS Service, Azure Web App, Google GKE Service)
- AWS Managed Airflow (MWAA) worker is counted as 1/2 unit. For a MWAA environment the number of workers are calculated as the average of minimum and maximum worker count
If a Kubernetes service is running with a Horizontal Pod AutoScaler (HPA), you cannot stop the service by clicking Stop in the service's Actions menu in the DuploCloud Portal.
Instead, do the following to stop the service from running:
- 1.In the DuploCloud Portal, navigate to DevOps -> Containers and select the Service you want to stop.
- 2.From the Actions menu, select Edit.
- 3.From the Replication Strategy list box, select Static Count.
- 4.In the Replicas field, enter 0 (zero).
- 5.Click Next to navigate to the Advanced Options page.
- 6.Click Update to update the service.
No. DuploCloud is calling the cloud provider's API directly. Based on user requirements, the software interacts with the cloud provider API asynchronously, maintaining a state machine of operations with built-in retries to ensure robustness. Any configuration drift, system faults, security, and compliance controls are monitored continuously by interacting with the cloud provider.
The Terraform and DuploCloud Web UIs are layered on top of the DuploCloud platform.
User Interaction with the DuploCloud Platform
DuploCloud provides an SDK into Terraform called the DuploCloud Terraform Provider. This SDK allows the user to configure their cloud infrastructure using DuploCloud constructs, rather than using lower-level cloud provider constructs. This enables the user to get the benefits of Infrastructure-as-Code, while significantly reducing the amount of code that is needed. The DuploCloud Terraform Provider calls DuploCloud APIs. Our DevOps white paper provides detailed examples.
You should create a separation between your developers and your DevOps team. Allowing developers to use the Web UI in non-production development environments, to quickly iterate product changes, can create an inconsistency.
In both production and critical non-production environments, the cloud services setup, as well as first-time application deployment, can be done via Terraform code. CI/CD workflows can be built that only update the application (Docker and Lambda) deployments. Any cloud service level change should be done by the DevOps team via Terraform and developers should ask the DevOps team to do the same. Developers should still be able to trigger CI/CD for their application rollouts without DevOps involvement.
CI/CD is the topmost layer of the DevOps stack. DuploCloud should be viewed as a deployment and monitoring solution that is invoked by your CI/CD pipelines, written with tools such as CircleCI, Jenkins, GitHub Actions, etc. You build images and push them to container registries without involving DuploCloud, but invoke DuploCloud to update the container image. An example of this is in the CI/CD section. DuploCloud offers its own CI/CD tool, as well.
DuploCloud's out-of-the-box diagnostics stack is optional. To integrate with a third-party toolset like Datadog, you follow the toolset's guidelines and deploy collector agents. You can do this as if you are running an application within the respective DuploCloud tenants.
Last modified 6d ago