Rootconf
Rootconf

Rows, columns, and consequences

Speak at Rootconf’s Special Edition on Databases

All submissions

Aditya Bhalla

Custom Autoscaler for using Azure Cosmos DB in a cost effective manner

Submitted Apr 30, 2026

Session type - select the format for your session: 15-minute talk – focused engineering experience

Abstract

Cloud-native databases like Cosmos DB promise elastic scalability, but in practice, throughput management often becomes a manual, reactive, and error-prone process. Teams tend to overprovision to avoid throttling—leading to wasted cost—or underprovision and risk latency spikes and SLA violations. This talk presents a proactive autoscaling system for Cosmos DB that eliminates this trade-off by dynamically adjusting throughput using real-time telemetry and historical workload patterns.

The session will dive into the engineering challenges of building a stable autoscaler—handling bursty and unpredictable traffic, avoiding scaling oscillations, respecting minimum RU constraints, and enabling low-latency decision-making. It will also cover how feedback loops and telemetry pipelines were designed to ensure reliable, production-grade scaling. The discussion is grounded in real-world learnings and demonstrates how databases can be made to behave predictably at scale while optimizing for both performance and cost.

Takeaways

  • A practical framework for building proactive autoscaling systems for distributed databases, including handling noisy traffic patterns, designing feedback loops, and preventing oscillations in scaling decisions.
  • Actionable strategies to optimize cost-performance trade-offs in Cosmos DB, such as dynamically right-sizing RUs, leveraging telemetry effectively, and avoiding common pitfalls in autoscale configurations.

Who Would Benefit

  • Backend engineers working with Cosmos DB or similar distributed databases
  • SREs and platform engineers focused on scaling, reliability, and cost optimization
  • Architects designing high-throughput, low-latency data platforms

Speaker Bio

Aditya is a senior engineer at InMobi, working on large-scale distributed systems for ad-tech platforms. He specializes in high-performance data infrastructure, with deep experience in Cosmos DB, Aerospike, Spark and real-time systems. His work focuses on building intelligent, cost-efficient platforms that operate reliably at scale.

All submissions

Comments

{{ gettext('Login to leave a comment') }}

{{ gettext('Post a comment…') }}
{{ gettext('New comment') }}
{{ formTitle }}

{{ errorMsg }}

{{ gettext('No comments posted yet') }}

Hosted by

Rootconf
Rootconf
We care about site reliability, cloud costs, security and data privacy