Treebo uses Facets to shift-left on infrastructure management and boosts Dev and Ops efficiency

About Treebo

Treebo Hotels started in 2015 and is one of India's most popular hotel chains in the premium-budget segment. With over 600 hotels across 100+ cities and over 2 million guests, they’ve used technology to transform the travel and hospitality industry.

Hotel Superhero is Treebo’s SaaS offering in the hospitality-tech space. It is an all-in-one integrated software for accommodation businesses of all sizes.

"DevOps is an important part of Treebo Tech, even more so for our SaaS product with global offerings and catering to international audiences. We need to optimize for performance and optimize infrastructure to build, deploy and monitor services in production." - Kadam, Co-Founder & CTO, Treebo.

The challenges

Dev and Ops Team Productivity: Treebo had in-house infrastructure automation and workflow tools for their build and deployments.However, the dev and ops teams found that a majority of their time was spent in the running and maintenance of these in-house tools.

In addition, resolving production issues took up valuable time, hampering overall productivity.

"The biggest challenge for the Ops team was that they would spend a good 70-80% of their time in solving production issues or helping the development teams out in debugging, this was frustrating for both teams. The turnaround times of the team would inevitably be higher because of this." - Kadam, Co-Founder & CTO, Treebo.

Multiple Production Environments: The productivity problem became worse when the new SaaS venture, Hotel Superhero, was introduced. It was built on Treebo's tech.Treebo is primarily an India-based business. This meant that the production environment was set up in only one region.

Architecting Hotel Superhero brought up challenges in replicating this infrastructure while maintaining consistency.

"We have multiple production clusters and it’s mission critical for the success of the business itself that each production cluster has identical specs." - Kadam, Co-Founder & CTO, Treebo.

From concerns to certainty

Treebo started a pilot  to mitigate two primary concerns:

• Would Facets limit self-learning for their DevOps Teams and abstract everything from them?

• Would they end up with a strong dependency on Facets, that would prevent their own teams from enhancing things?

Facets do not replace DevOps teams. By automating repeated and redundant tasks, the platform gives DevOps teams the time to focus on business-relevant problems. Facets is built on well-known technologies and provides transparency on how they are used. Further, it is extensible, encouraging DevOps teams to build on top of it.

As Treebo began to evaluate Facets, they were pleased that it made life simpler for their teams.

"As we built familiarity with Facets, the concerns were mitigated. Since Facets is built over standard offerings like Kubernetes, we can still build much more on top of it! The combination works really well. Facets is only making it simpler which otherwise would need 3X more time and resources. With instances of co-architecting and problem-solving, our team developed faith in the product and the team." - Kadam, Co-Founder & CTO, Treebo.

How Facets helped Treebo

After the evaluation period, Treebo migrated all their services, databases, queues, and caches from their own management layer to Facets. They rolled out the new setup on the Staging environment and then migrated to Production. With this move the advantages they gained were:

• Drift-free environments with a single source of truth.

• Ability to spin up and terminate environments with a single click.

• UI-driven workflows for SRE and dev teams on alerting, logging and monitoring

• Template-based tenant-specific Infrastructure: Treebo had a requirement that needed certain dedicated resources to be created dynamically when a new tenant is onboarded. This feature was developed by Facets where such resources get templated, radically improving the manageability of the system.

"With Facets, we have been able to set up true infrastructure-as-code including alerting, monitoring, deployment, and infrastructure creation…So our Staging environments now look very identical to Production environments and all production clusters which in turn look similar, this is a big win for us. So in case of production issues, one thing that our teams can be sure of, is that there are no infrastructure differences. It narrowed it down to only the Code issues."- Kadam, Co-Founder & CTO, Treebo.

Key results

• Stable Environments - 70% Lesser Production issues with consistent infrastructure.

• Faster Tenant Onboarding - From Days to Minutes.

• Simplified pipelines with CI and CD decoupled - 50% reduction in time taken for Deployments.

• Empowered DevOps Team - 80% of DevOps team time is now spent on making new tooling, guidelines, and research to provide a better experience to developers.

• Shift-Left of infrastructure management - Finally, Treebo had a system where developers see through how every single change gets propagated to the environment as they continuously evolve the architecture. This led to an organic movement towards shift-left infrastructure management without shifting left the complexities associated with it.

"Earlier the development teams were only worried about their code. Facets have created awareness for Developers - I can already see developers now aware of, How will my code be deployed? What kind of infrastructure does it need? What kind of alerting needs to be set up? What kind of autoscaling is needed?"- Kadam, Co-Founder & CTO, Treebo.