Uber Advanced Technologies Group (ATG) is part of the pace-setting American multinational ride-hailing company. Its mission is to introduce self-driving technology to the worldwide Uber network and transport people and goods more safely, efficiently, and affordably.
In early 2021, Uber ATG was acquired by Aurora, an autonomous vehicle company.
You may be interested in the on-demand webinar, How Uber ATG Reduced AWS Costs 15% in 30 Days
Customer Challenges / The Opportunity
Manage dynamic change in the AWS environment without slowing Uber ATG’s high-velocity teams. Get actionable insights on infrastructure changes that affect costs, security, performance, reliability, and operational excellence.
Fast-growth Uber ATG was running extensive workloads on AWS, with continual changes to their environment. The Uber ATG team surveyed the market’s existing tools, always striving to improve and keep up with innovation. They found that the current tools were not built for modern and dynamic cloud infrastructures with serverless and containers. Those tools lacked sufficient anomaly detection and notification and had static pdf reporting. So, the Uber ATG teams created a proprietary tool that provided some of the functionality they wanted. However, this created a problem for the Uber ATG teams because it was time-intensive, consuming valuable engineering resources.
While creating their proprietary tool and reviewing current market technologies, the Uber ATG team learned about nOps, a SaaS cloud management platform. They engaged nOps for an extensive three-month technical evaluation. nOps provided proof-of-concept (PoC) evaluation, new feature development, and live trial simulations.
Following the PoC, Uber ATG selected nOps as its cloud-native management platform to operationalize costs, provide reports for leadership and budget processes, and improve performance efficiency. nOps was tasked to consolidate the Uber ATG architecture, including the migration of earlier reports and integration with Finance.
nOps monitors changes across Uber’s AWS environment, providing detailed alerts on changes in cost, threats to security, functional reliability, performance efficiency, and operational excellence. nOps’s alerts include detailed context to improve response time and pinpoint specific instances for managing necessary changes.
The decision to use nOps rather than continue building a proprietary tool was a double savings for Uber ATG. Not only did it save 15% in AWS costs in 30 days, followed by additional savings, but it also freed up valuable engineering time. This time savings can be devoted to delivering new product features faster.
Modern change management aims to strike the right balance of automation and just-enough lightweight process for control without compromising DevOps speed. For example, using nOps, Uber ATG received detailed, automated alerts on anomalies in their AWS infrastructure that provided the ability to perform root cause analysis (RCA) at the resource level and effect repairs more quickly.
The team at Uber ATG viewed its AWS infrastructure as an enabler of business speed and agility. They leveraged nOps to:
- Enable seamless governance so that departments could gain budget approvals and demonstrate ROI without disruption.
- Deliver cloud management 2.0 and reimagine approval processes.
- Do smart budgeting based on nOps’s machine learning-based detection of billing anomalies (costs in the infrastructure that do not fit typical parameters). For example, nOps identified that Uber ATG’s Amazon Athena costs had grown during a several-month period by 5000%! Based on this insight, Uber ATG took action to work on more efficient usage and integrations to enable more team members to take advantage of Athena internally.
- Implement chargebacks — nOps enabled Uber ATG to allocate cloud infrastructure costs easily and accurately across the organization at a level of detail they had only imagined. Here are some examples:
- Where the organization was sharing a large Kubernetes cluster (typical of many enterprises), they could charge by service or deployment based on a label in the Kubernetes cluster. 15% AWS COST SAVINGS IN THE FIRST 30 DAYS ENABLED AUTOMATIC CHARGEBACKS BY DEPARTMENT FOR AWS COSTS FACILITATED ROOT CAUSE ANALYSIS OF AWS INFRASTRUCTURE ANOMALIES AT THE RESOURCE LEVEL
- They could automatically charge back a department for a user’s Athena query or a Glue (serverless Spark) job – where these workloads were common and generated hourly costs.
The nOps and Uber relationship demonstrates the most fundamental and primary principle upon which nOps prides itself. That is, in every relationship, the needs, welfare, and success of the customer are the driving force behind nOps. Cost savings are only one component of a good business strategy. Not only did nOps help Uber ATG save money, but nOps also empowered the Uber ATG team with more productive use of their time, which is often an unquantified benefit. nOps gives its customers broad-spectrum savings that encompass more than saving overhead costs.