How Kirana11 Scaled AWS RDS MySQL & Cut Database Costs by 50%

Overview

Kirana11, an Indian online grocery platform, faced severe database crashes and checkout disruptions during sudden promotional traffic surges. Partnering with Mydbops, they implemented targeted query tuning and optimized resource sizing. This intervention successfully stabilized their cloud database, eliminating service interruptions during peak events, restoring rapid application response times, and significantly lowering recurring hosting expenses to protect their overall operating margins.

↑

↓

50%

Annual Cost Savings

By shrinking the AWS RDS instance size by half.

↑

20K

↓

QPS Handled

Scaled query capacity from under 7,000 to 20,000 QPS.

↑

↓

CPU Load Drop

Reduced database resource utilization fivefold.

Zero

↑

↓

Database Crashes

Eliminated high-latency bottlenecks during promotional launches.

MySQL

Consulting Services

About

Kirana11 is an online retail platform designed to help local kirana stores reconnect with their customers, expand sales opportunities, and build long-term relationships. Operating via both a mobile app and web portal, Kirana11 serves as a digital bridge in India's grocery e-commerce landscape.

★★★★★

Mydbops helped us smoothen our DB within short span of time, highly appreciate their support

Krishna

Tech lead, Kirana11

Deployment Type

Database Stack

Outcome

Cloud-Based Deployment

RDS MySQL

50% Instance Cost Reduction & 5X CPU Load Decrease

Deployment Type

Cloud-Based Deployment

Database Stack

RDS MySQL

Outcome

50% Instance Cost Reduction & 5X CPU Load Decrease

Business Challenges

Overview

Following the launch of attractive consumer offers, Kirana11 experienced a sudden surge in traffic. While these promotional campaigns succeeded in driving consumer interest, they created severe issues for the platform's backend:

Slow Customer Transactions: Database writes and updates were slow, resulting in delayed checkouts and sluggish order processing.

Increased Slowness & Latency: High query volume caused massive latency spikes on the mobile application and website, causing friction in the buyer journey.

Platform Downtime: The database was unable to handle workloads exceeding 7,000 queries, leading to regular database crashes during high-traffic promotional hours.

Operational Risk:The team required proactive database support during marketing launches to prevent revenue loss and protect customer relationships.

Goals

The key objectives the client was aiming to achieve:

→ Eliminate database crashes and stabilize the platform during peak promotional windows.

→ Scale the query capacity of RDS MySQL to comfortably handle traffic spikes.

→ Reduce database read/write latency to improve the front-end user experience.

→ Optimize database resource usage to control escalating cloud infrastructure costs.

Risks if Not Addressed

If left unresolved, these challenges posed serious risks

Risks & Impact if Not Addressed

Performance Issues

Without resolving replication lag and fragmented tables, query performance would continue to degrade, leading to a frustrating customer experience during peak hours.

Business Continuity Risks

Non-standardized backup policies increased the risk of data loss and prolonged outages, potentially disrupting thousands of orders in real-time.

Revenue Loss

Poor performance and downtime during peak times directly impacted Swiggy’s ability to fulfill customer demand, resulting in lost revenue and dissatisfied users.

Escalating Costs

Continued reliance on oversized, under-optimized infrastructure would lead to unnecessary monthly spend, straining the company’s profitability.

Developer Inefficiency

Lack of a stable and scalable database foundation meant developers spent significant time firefighting performance issues instead of innovating on features.

Performance Issues: Replication lag and fragmentation slow order searches and transactions.
Business Continuity Risks: Non-standardized backups mean longer recovery times and higher data-loss risk.
Revenue Loss: Slow page loads or timeouts during peak hours lead to failed checkouts.
Escalating Costs: Over provisioned, under-optimized servers strain profitability
Developer Inefficiency: Engineers spend more time firefighting than building new features

Goals

The key objectives the client was aiming to achieve:

→

[Goal 1]

→

[Goal 1]

→

[Goal 1]

Solution Provided by Mydbops

To address the live production issues immediately, Mydbops deployed consulting specialists directly to Kirana11's Bengaluru headquarters under Emergency Consulting services:

▸ Database Bottleneck Audit:

Conducted a detailed analysis of the active RDS MySQL environment to identify exact performance blockages under load.

Operational Throughput Capacity

Query capacity comparison during peak launch events

Pre-Optimization Limit

7,000 QPS

Frequent System Stalls

High read/write latencies resulted in slow application load times and frequent database crashes during launch events.

Optimized Stable Limit

20,000 QPS

Stable High-Throughput

Successfully handles peak launch periods without lag or service interruption, ensuring stable store transaction checkouts.

‍

▸ Resource Sizing Optimization:

Re-evaluated resource consumption patterns to align database size with actual transactional needs.

‍

▸ Targeted Query Optimization:

Restructured and tuned slow-running queries to reduce response times and lower the overall processing burden on the system.

Database Routing & Optimization Flow

Data transaction pathway following database sizing and query optimization

Client Tier

Customer Mobile Apps & Web Portals

→

Application Tier

Scalable Production Web Servers

→

Optimized DB Layer

AWS RDS MySQL (Tuned Query Routines)

→

Value Generation

50% Database Cost Savings

Results and Impact

Key Outcomes

✅ Substantial Cost Savings

By optimizing queries and resource consumption, Kirana11 was able to reduce its active database instance size by 50%. This change immediately halved their monthly AWS database hosting spend.

‍

✅ Uninterrupted Promotional Campaigns

The database successfully handled traffic spikes of up to 20,000 QPS nearly three times the previous limit—allowing marketing teams to run high-volume sales campaigns without system interruption.

‍

✅ Improved User Experience

Resolving high write latency and slow response times restored smooth app and website performance, preventing cart abandonment and securing buyer trust.

‍

✅ Efficient Resource Use

CPU utilization on Amazon RDS dropped by 5x. This reduction freed up significant computing headroom, ensuring stable database operations during peak shopping hours.

CPU Workload Consumption

Maximum CPU utilization on Amazon RDS during promotional events (Lower is better)

Pre-Optimization CPU Peak (Unstable Environment)

100% Workload (Extreme Crash Risks)

Post-Optimization CPU Peak (Mydbops Managed Environment)

20% Workload (Safe Operating Headroom)

0% Utilization 5X Drop in Compute Resource Demand 100% Resource Exhaustion

Operational Impact & Metrics Matrix

Comparison of system configurations, resource efficiency, and organizational outcomes

Performance Indicator	Pre-Optimization Profile	Optimized Profile	Net Business Outcome
Peak Transaction Capacity	Under 7,000 QPS	20,000 QPS	Comfortably holds transactional load during sales events
RDS Compute Footprint	Exceeds 100% Load	20% Peak Load (5X Improvement)	Maintains stable headroom without performance degradation
Database Cost Structure	Over-provisioned Instance	50% Smaller Instance Size	Decreased ongoing infrastructure overheads
Service Reliability	Prone to DB Crashes	Stable Zero-Crash Performance	Ensures a dependable checkout journey for customers

When Kirana11 launched new sales promotions to help local grocery stores reach digital buyers, customer demand quickly tested the limits of their infrastructure. At 7,000 queries, database bottlenecks triggered system slowness and unexpected crashes. This operational friction threatened to impact both the customer experience and merchant relations

‍

By deploying database experts directly to Kirana11's headquarters, the team isolated the underlying query and configuration issues. The resulting performance improvements did more than just stabilize the platform; they allowed Kirana11 to handle nearly three times its previous traffic peak while cutting database hosting costs in half. What began as an urgent system bottleneck became an opportunity to establish a highly efficient, cost-optimized data layer ready for future business growth.

‍

If your growing platform is experiencing database bottlenecks, sluggish transaction speeds, or rising cloud hosting overheads, our database consultants are here to help. Reach out to our team today to discuss how we can stabilize, optimize, and secure your data operations.

Consult with a Certified Database Expert

Consult with a Certified Database Expert →

Need Expert Database Solutions?

Talk to a Database Expert Today!

Database solutions are provided by mydbops expert team

Mydbops set up High Availability (HA) Solutions with InnoDB or Percona Clusters, ensuring continuous uptime and fault tolerance.

Thank You!

We’ve got your request, our expert team will be contacting you shortly.

Oops! Something went wrong while submitting the form.