Table of content
Who are we?
NT-ware Operations is a global team responsible for the smooth operation of uniFLOW Online. We have team members in Europe, the APAC region, and the USA, including support specialists, Azure infrastructure specialists, and developers (DevOps), amongst others, carrying out a wide range of activities, both proactive and reactive. Additionally, this team is responsible for building, maintaining, and expanding the infrastructure that supports uniFLOW Online.
What we do?
uniFLOW Online is monitored 24/7 using industry-leading tools on the Azure platform and the Elastic log analytics engine. NT-ware Operations uses these tools, together with other monitoring systems, to aggregate data and conduct analysis to diagnose, dissect, and preempt situations, allowing a fast response to alerts and providing high-quality feedback into our "development and improvement lifecycle".
The responsibilities of the NT-ware Operations team go well beyond monitoring uniFLOW Online. The team is tasked with all deployment-based activities for uniFLOW Online as-a-service and the supporting Microsoft Azure infrastructure. We work directly with other NT-ware departments, acting as an escalation point for issues raised by Support, Development, and Architecture teams to ensure the growth of uniFLOW Online.
With such a large group spread across the globe, we leverage industry-leading tools to support our communication, ticketing, incident management, and documentation needs. These systems must be highly secure and highly available. To achieve this, NT-ware uses Microsoft Teams and other Microsoft communication platforms to maximum advantage. To support our ticketing and documentation requirements, NT-ware uses Atlassian Jira Software and Confluence in the cloud.
How do we handle incidents?
No one wants to see an incident occur, but it will happen. How quickly we respond and learn from them is key. The NT-ware Operations team is responsible for all incident management and follows a defined incident response process. This is built on industry best practices and many years of field support experience:
- Monitoring and/or an alert will trigger an incident response plan to start. This involves identifying the incident, its impact, and, finally, the scope.
- The communication plan will start in parallel, ensuring timely and transparent information exchange.
- NT-ware Operations will coordinate with Development, managing escalation or system changes to respond to the incident.
- All actions and communication are coordinated through our structured Microsoft Teams channels.
- To ensure our Service Level Agreements (SLAs) are monitored and measurable, we will also create a ticket in our issue tracking system.
- This is a vital step to ensure we have a reference point for all corrective actions.
- We have a strict review and "Lessons Learnt" process required for any incident prior to it being considered closed.
More information here:
What is our development and improvement lifecycle?
Developing a service on the scale of uniFLOW Online is a big undertaking. Before we deploy new features and improvements, they undergo rigorous internal testing. The development and improvement lifecycle is built into our Jira development pipeline with checks at every critical stage. The NT-ware Operations team follows the product lifecycle to ensure monitoring and performance metrics are fed back into the "continuous development" process.
NT-ware Development follows the Feature-Driven Development (FDD) process in four phases, which enables teams to update the project regularly and identify errors and potential problems quickly.
The four phases can be described as:
- Initial feature description.
- Creation of a feature list that resembles the overall model.
- Defining actual development tasks per feature.
- Final feature development deployment.
Safe deployment practices
NT-ware deploys updates to the service through a process known as the "Ring" deployment approach. The general idea is simple: a given release should be disclosed initially to only a small group of users with the highest tolerance for risk. Then, if the release is working as expected, it can be released to a broader group of users. If it is still on track, the process can continue to wider groups of users or in stages. The following stages form part of our development process:
- Hourly and nightly uniFLOW Online deployments take place on "internal deployments" during the development phase.
- These deployments are continuously subjected to our QA automation and testing systems.
- Real customer data is never utilized for testing or other purposes in any pre-production stage of uniFLOW Online.
- To progress towards the production stage, we have further uniFLOW Online deployments that see feature and product testing by "Hands On" teams consisting of NT-ware and Canon Quality Assurance departments. These releases form part of our Early Access (EA) and BETA and deployments to ensure the service and features have been thoroughly tested before release into the production deployments of uniFLOW Online across the globe.
- At the EA stage, NT-ware utilizes uniFLOW Online in our headquarters and remote offices. This "dogfooding" is a vital step in bringing our service and capabilities to our staff.
uniFLOW Online deployment updates/upgrades
uniFLOW Online deployment updates/upgrades are carried out without any downtime or impact on customers and end users. Microsoft Azure allows deploying services and components to a staging slot next to the production slot for uniFLOW Online. This includes the uniFLOW Online application and configuration elements. Having everything ready within the staging slot ensures that all functionalities and scaling are available before being transferred to production. Traffic redirection is seamless, and no requests are lost during the transfer operation.