Search…
Vamp Concepts
Release Agent
Integrations
FAQ
All the answers you want to know about Vamp Cloud Continuous Release Orchestration
What is Release Orchestration?
Release Orchestration is the management, direction, and automation of a software release (or Go-Live) process. All the steps, checks, gates, and validations involved in a typical software release can be managed by a Release Orchestration system like Vamp. Steps can include a certain percentage of visitors, and/or a specific visitor-segment using OSI Layer 7 elements, such as geolocation, device type, headers, cookies, or IP addresses.
Multiple steps can be chained. For each step, a Release Orchestration system validates for both technical and business metrics to establish whether the release is still within an acceptable quality range. These ranges are typically defined by a combination of DevOps and SRE “golden metrics”, thresholds, and error budgets, SLA/SLO-based targets, and ML-powered anomaly detection. There can also be rolled-up top-line metrics, like Vamp Health, that give an “early warning signal” when the quality of a release decreases.
A Release Orchestration system can orchestrate releases within environments (intra-cluster), and also over different environments and clusters (inter-cluster). A Release Orchestration system typically uses traffic-shaping mechanisms like proxies, load balancers, Ingresses, and service meshes for its segmentation.
How is Release Orchestration different from Continuous Delivery or Continuous Deployment?
Continuous Delivery or Deployment ensures runtime artifacts (such as containerized services and applications) from a CI pipeline are deployed to a specific environment and are running correctly. Releasing is the process of “going live” with the new software version in a controlled and safe way. It is focussed on the business and the end users.
Is Vamp a CI/CD tool?
Vamp is a Release Orchestration system. It extends and enriches your current CI/CD solution with automated and smart features to achieve a controlled, safe, and scalable process for your release.
Does Vamp only work with Kubernetes?
No. Vamp is cloud native and tightly integrated with Kubernetes and Kubernetes-related cloud-native components. However, Vamp's release automation also works with non-Kubernetes traffic-shaping mechanisms, such as HAProxy and NGINX, and cloud vendor load balancers, including AWS ALB. Vamp’s event and metric detection, validation, and Machine Learning (ML) processes support non-Kubernetes metric stores and data sources. Vamp also works with other systems, such as AWS CloudWatch, Google Stackdriver, event streams (for example, Kafka, AWS SQS, or NATS), Google Analytics, and Sentry.io.
Isn't Canary releasing handled by a service mesh, such as Istio?
You can configure a service mesh to use either static routing or traffic splitting. However, the service mesh does not know when it is safe to proceed to the next release step, or what this next step should be. In other words, a service mesh is just an abstraction layer on top of proxies.
A Release Orchestration system like Vamp is a smart engine that both uses and drives a service mesh (and/or other components like Ingress and external load balancers) by providing it with dynamic configuration and continuously observing the effects and results of that configuration.
How does Vamp relate to feature flags and toggles?
Feature flags and toggles are useful, allowing you to develop and test features without exposing them to your audience, or releasing them only to specific customer segments.
Canary Releasing, which is an important pattern and feature in Release Orchestration systems like Vamp, creates a safe, fast, and automated process for going live with your new software. We believe that Layer 4 and Layer 7 of the OSI model provide a better solution for creating traffic splits, conditional routing, and dynamic releasing. You do not need to instrument your code, there is no code overhead of adding additional logic, and the Layer 4 and Layer 7 mechanisms already exist.
How much impact will Vamp have on the resource and latency of my application?
Vamp does not add any components inside your application's end-to-end flow, it simply dynamically configures the existing Layer 4 and Layer 7 traffic-shaping mechanisms. More complex rules can add negligible overhead, but solutions like HAProxy, NGINX and Envoy are highly efficient.
How secure is Vamp?
Vamp is very secure. We use best practices such as mTLS, RBAC, Auth0 and HashiCorp Vault time-based tokens. Only rolled-up and sanitized, high-level metrics are passed from the client infrastructure to our central management environment. All data is isolated and encrypted. Enterprise needs are recognized and reflected in the design of Vamp.
Do I need a service mesh for Vamp to work?
No. Vamp can also work with plain Ingress solutions, such as NGINX and Contour (Envoy-based). Vamp’s release agent checks whether the selected Ingress solution is already installed and running. If not, the agent automatically installs and configures it for you.
What does Vamp cost?
See our pricing page www.vamp.io/pricing. Contact us for special discounts for open source, community, and non-profit projects.
Can we have a better service level with our current product tier?
Last modified 2mo ago
Copy link
Contents
What is Release Orchestration?
How is Release Orchestration different from Continuous Delivery or Continuous Deployment?
Is Vamp a CI/CD tool?
Does Vamp only work with Kubernetes?
Isn't Canary releasing handled by a service mesh, such as Istio?
How does Vamp relate to feature flags and toggles?
How much impact will Vamp have on the resource and latency of my application?
How secure is Vamp?
Do I need a service mesh for Vamp to work?
What does Vamp cost?
Can we have a better service level with our current product tier?