Development Operations (Dev Ops)
ONQU will improve your development and delivery capability by using DevOps with fully defined metrics and monitoring
Great advances have been made in application development and the delivery of support services to organisations operating Agile delivery frameworks.
This has been encapsulated with the introduction of Development Operations ‘DevOps’ and the implementation of refined working practices. Implemented correctly this removes cost barriers, improve quality and significantly increases delivery velocity and capability whether deploying software on premise or in the cloud.
Clouds on the Horizon?
However, there are already some recognised issues with the traditional DevOps approach. DevOps now must operate in an ever-increasing complex technical eco-system that is delivered at an accelerated tempo using Agile methodologies.
Technical debt and the rise of multiple interfaces, micro services and cloud deployments is adding infrastructural complexity and is increasing the number of dependent systems a deployment relies on and interacts with.
When deploying development to a modern software stack, the issue is no longer one of speed, but instead is one of trust.
Organisations are now asking these questions:
Do we trust what has been delivered?
• Is it the right version?
• Is it deployed as we expect?
• Is it deployed on all the nodes/systems in our architecture framework?
Do we trust that the platform is working as expected?
• Has a deployment introduced issues?
• Is every component on every node in a healthy state?
• Are we sure we haven’t introduced security issues?
Do we trust we know when things have failed?
• Is there accurate monitoring of service health?
• Can it account for failed deployments on some but not all nodes?
• Does it surface issues in a meaningful/actionable way?
These questions need to be answered to assess how an application deployment is working and whether it is meeting the objectives of the business. To create this type of evidence based service, DevOps needs to be measured against a set of defined metrics.
Metric Based Development Operations
Metrics and measurement is now defining a new working framework that brings together development, quality assurance and test engineering in a unified delivery process.
The key approach is to map out the delivery stream and then create key performance indicators (KPI’s) and specific metrics that are then used to measure the deployment with.
Once defined, these metrics are monitored to identify deployment problems, release trends and anticipate future states. This allows the development operations to fine tune the release processes to improve delivery and reduce failures in the form of expensive roll backs or bug fixing.
Evidence Based Operations
Metrics are the lifeblood of monitoring, and by utilising the best in class open source time series and monitoring applications a wealth of monitoring data becomes available.
Once defined, these metrics identify deployment problems, release trends and anticipate future states. This allows the development operations to fine tune the release processes to monitor point in-time failures as well as introducing trend analysis along with, before and after reporting. This significantly improves delivery and reduces failures in the form of expensive roll backs and bug fixing.
This approach has already seen great success when used large online, media and finance companies like Hotels.com, Sky TV and Universal Music UK. It’s a working practice that powerfully enhances DevOps and puts monitoring front and centre of all the stakeholders within the application life cycle. This encourages metrics and monitoring not to be just an add on to the current workflow, but an integral and important part of delivering reliable, secure and performant software.
Metric Centric DevOps Helps Who?
The Metric DevOps structure helps the following functions to include:
Developers – gain a clear insight into the performance and interactions of their software with an understanding of how issues will be seen by QA and support who don’t have access to debugging tools.
Quality Assurance – Have out of the box tools that can explore, visualise and validate individual components as well as the whole system. Including reproducing failure scenarios and wargaming playbooks with support.
Support – Having early access to learn and validate support tools to enable them to visualise, report and run recovery scenarios
Business – Business users have early confirmation that the system integrates as a whole and is fit for purpose, delivering business value. Ensures that failures are detected easily and quickly so owners are rapidly identified. The metrics, reporting and monitoring, provide for collective future planning to ensure that the systems support business growth.
Want to Know More?
ONQU can help you reach your Software delivery destination with
1. DevOps metric brief
2. Show and Tell Workshop
3. Proof of Concept Delivery
4. Coaching and Knowledge Transfer
If you feel that Metric DevOps can deliver a positive change in your development environment and you would like to discuss how ONQU can help to implement the changes you require please contact us via the form below.