How Inshorts ensured 100% Observability on GCP

Inshorts is Asia’s most-read news app with more than 10 million users and is making big strides with its byte-sized news portal. Inshorts strikes a chord with millennials who don’t have enough time to read whole articles but want to be updated with news and this is where their app comes in place with summarized news bytes that hardly take a minute to read.

Industry

Digital Media

Type

For Profit

Location

India

Challenge

With an increasing user base, they were also experiencing increased Google Cloud Platform bills and it was becoming a major chunk of their operating costs. The engineering team at Inshorts tried to analyze their cloud bills using GCP’s Billing Console but its limited visibility was holding them back.

We were using GCP Billing earlier and our major pain point was the delay in our metrics which at times were delayed by 24 hours. The real-time data that we get from OpsLyft has helped us curb many instant surges in cloud costs.

Manish Bisht

Head Of Technology, Inshorts

Challenge

Manish and his team went through their bills and reports and found out that a major portion of their cost bills went un-attributed on the GCP console. The reason behind this is the lack of granular visibility provided by the GCP console that makes it difficult to find even the cost of compute engines. For this, an efficient tagging system was needed.

The engineering team at Inshorts needed to get hold of their GCP bills as with an expanding user base their prime concern was a mix of providing new features and a robust platform to their users. Manish and his team would spend upwards of 8 hours a week trying to figure out their cloud bills by exploring the GCP console and tweaking every part of it but still couldn’t figure out how to map the unattributed costs.

Solution

When Manish and his team tried OpsLyft they discovered that Cost360 not only analyzes the cloud resources in real-time to give software developers insights and suggestive actions for making cost-friendly infrastructure changes but also discovers the unattributed resources in real-time to the most granular level and finds the amount of cost on those resources that was going untracked. The team at OpsLyft worked with Inshorts and curated a robust tagging strategy that will logically group the GCP infrastructure by applications, clusters, team, deployment, etc.

In addition to the unattributed costs, for the tagged resources, Cost360 showed them the exact resource id and other granular details which helped them identify which resources were the major cost contributors.

After tackling the problem of granular visibility, Manish & team were able to use the platform to shut down the unused and idle resources in their GCP infrastructure. This resulted in massive cost savings as well.

Also, we found out that Manish and team were evaluating APM solutions for efficient infrastructure and application monitoring. Our team is currently working with them to extend the platform capability to provide system usage monitoring of all GCP resources and the performance metrics of applications hosted on them within the same dashboards where they can see the cost observability as of now. With these additional solutions in place, Inshorts will achieve a full stack observability through a single platform, which otherwise they would have used 3 different products to do so.

Results

By leveraging Cost 360 and the services of the OpsLyft support team, understood the major contributors of their cloud bills. They were able to get granular visibility of their whole Google Cloud infrastructure, giving them a clear idea of how to manage their cloud costs. Manish and his team:

• Discovered untracked costs for about $400,000 after achieving a 100% coverage of resources across production and non-production environments through Cost360
• Implemented infrastructure changes worth $200,000 through actionable insights on the platform that were used by developers to investigate cost optimization issues without spending their precious bandwidth on analyzing information on GCP Billing console and finding patterns around the same.