MPL SRE Team Achieves Faster Cloud Migration and 8X Releases with Facets

With Facets as their enabler, MPL's SRE team achieved zero-downtime GCP migration, 8X faster releases, and self-service for 150+ developers.

Strategic Value Delivered by Facets

Situation

Mobile Premier League (MPL) is the world’s leading skill gaming platform with over 120 million users and 150+ engineers. During peak gaming seasons like IPL, traffic can surge 3X and an hour of downtime can lead to Millions of Dollars of loss in GMV. Every service needs to stay online—every release has to be smooth.

The migration also needed to account for critical workloads where an hour of downtime could lead to millions of dollars in losses, and handle peak traffic events like the Indian Premier League (IPL) which could increase request per second (RPS) by three times in key markets.

Initial State

1. Explosion of tools & delivery mechanisms: The software delivery setup involved a variety of tools like Harness, IaC, Custom Scripts, etc., which were introduced over time. This resulted in an explosion of pipelines that required effort to operate and maintain.

2. Limited Team Autonomy: Releases involved handoffs between Dev, QA, and SRE teams. The SRE team was maintaining pipelines instead of improving systems.

3. AWS → GCP Migration: A looming AWS → GCP migration, millions of dollars worth of cloudprint, had to happen without any downtime or freezing regular feature releases.

“If we could take the cloud migration as an opportunity to make our delivery pipeline self-serve and governed, that would yield multi-fold ROI on the effort we spend on the migration,” stated Kaushal.

"

The team needed a better operating model—one that empowers developers without sacrificing control and enables a safe migration.

Building MPL's Internal Platform using Facets

Rather than expanding scripts or layering on more automation, MPL's platform team took a new approach: model infrastructure and delivery as high-level intents, then orchestrate outcomes declaratively using Facets.

The platform team defined the architecture, policies, and patterns:

• What an application should declare for which they were earlier dependent on SRE teams (via intents)
• Which cloud and environment should it target (AWS/GCP, Dev/QA/Prod)
• How identity, networking, and observability should be wired in
• Which compliance and rollout policies apply

This internal platform, powered by Facets, became MPL's single source of truth for operating infrastructure safely at scale.

Phase 1: Service-by-Service Migration to GCP

Instead of freezing dev work for a big-bang migration, the platform team embedded cloud selection as a configurable flavor in each service's definition.

Developers could simply update the configuration to switch service from AWS to GCP or back—The Platform with Facets Orchestrator took care of the rest:

• Routing and upstream connections were handled automatically.
• Service health and rollback safety were built into the flow.
• Progressive rollout from Dev to QA to Prod.

We flipped a configuration, triggered a release, and the service was up on GCP—without touching the pipelines,—Said Kaushal.

"

This approach enabled the development team to

• Migrate at their own pace without compromising feature releases
• Switch over Service-by-Service without taking a cut-over risk
• Establish a Repeatable switchover and rollback process

Phase 2: Empowering Developers with Controlled Self-Service

Once migration was underway, the platform team focused on unlocking developer velocity without compromising governance.

With Facets, they exposed platform capabilities as declarative intents—each with guardrails baked in. QA and development teams could now:

• Deploy services without needing platform intervention
• Trigger progressive rollouts with safety nets
• View deployment health, traceability, and compliance status—all from one place

We didn't just speed things up. We gave teams ownership over their workflows, said Kaushal. That's when everything clicked.

"

This led to a dramatic increase in deployment throughput: from 1 to 8 releases per week.

Phase 3: Unifying Orchestration Across 100+ Services

Before Facets, different pipelines were used for provisioning, deploying, and configuring services. This caused inconsistencies, drift, and confusion.

The platform team modeled a cohesive orchestration layer with Facets that standardized how services and dependencies were defined and deployed across tools.

Facets became the orchestrator behind the scenes, coordinating:

• App deployments, Canary / Blue-green Rollout policies (Argo, Istio)
• Cloud infra provisioning (GKE, IAM, DNS, etc.)
• Secrets management
• Compliance controls

Now, every service flows through the same orchestrator, with consistent outputs and traceable workflows.

Environments as a Commodity

One major unlock was the way environments were modeled as a separate dimensions—not bundled inside the resource definition.

This allowed MPL teams to:

• Launch production-like performance environments with a click
• Tear down environments when unused—saving 20% on compute
• Maintain environment-specific policy mappings and access controls

The separation of intent (what the service needs) and environment (where it should run) gave MPL enormous flexibility and scalability.

Built-In Governance & Policy Enforcement

Security, identity, and compliance were not layered on—they were baked in. IaC modules were secured before being used and the platform team defined standardized policies and enforcement logic. Facets ensured these were applied consistently, across environments.

• Secure-by-design networking and IAM
• RBAC modeled per environment applied globally
• Infra drift detection and policy compliance checks

We have a lot of granularity in our RBAC, controlling every action and entity. A few days ago, someone performed a full release that he wasn't supposed to. We were able to remove just the full release privilege while leaving developers free to work - Vandan Rohatgi, Security Engineer, MPL.

"

What Made The Difference?

The MPL story isn't just about tooling—it is about platform leadership.

By adopting Facets, the platform team didn't just improve deployments—they built an internal cloud platform that developers trust. A system where:

• Teams declare intents and choose environments
• The orchestrator builds and operates everything in between
• The policy is never optional but never a burden either

They made Facets their platform engine, and in doing so, they gave developers superpowers.

"We now have a single source of truth for our Infrastructure. Also, we have enabled smarter auto scaling and faster environment replication because everything is working on Facets already.” - Kaushal Bagtharia, AVP of Engineering, MPL

"