下载PDF

实例探究 > How Netmeds switched from Elasticache to Redis Enterprise and achieved zero downtime

How Netmeds switched from Elasticache to Redis Enterprise and achieved zero downtime

技术

平台即服务 (PaaS) - 数据管理平台
应用基础设施与中间件 - 数据交换与集成

适用行业

电子商务
药品

适用功能

商业运营
销售与市场营销

用例

数字孪生
预测性维护
远程资产管理

服务

系统集成
软件设计与工程服务

挑战

With just three weeks to go before a big marketing event designed to deliver a significant increase in user activity, Netmeds was desperate to resolve the failure issues that were currently plaguing its Amazon ElastiCache. Huge spikes in user traffic were choking Redis with too many connections, causing it to fail. As a result, Netmeds would experience downtimes of up to an hour that easily lost them 750 orders every 30 minutes. Netmeds initially thought the issue could be resolved through clustering on ElastiCache, but the company’s PHP platform did not support clustering of Redis. Next, NetMeds evaluated connection management solutions as a way to reduce the number of open database connections. They looked at Twemproxy, Dynomite, and HAProxy in an attempt to set up connection pooling, but none of these solutions supported every command that Redis supported.

关于客户

Netmeds is an online pharmacy that ships products to every corner of the Indian market. The pharmaceutical’s omni-channel sales approach supports product orders via its website, mobile app, and call center. As a three-year-old company, Netmeds has been ramping up its marketing efforts in the very new space of pharmaceutical e-commerce. This has resulted in traffic spikes that caused its AWS ElastiCache instance to fail. Netmeds needed a solution that could handle the high availability and performance required by its hundreds of thousands of daily users.

解决方案

A simple Google search brought Redis Labs to Netmeds’ attention. Pandit reached out immediately and received a response from Redis Labs the very next morning. With less than three weeks to go live with a resolution before their big marketing push, it was a huge relief to learn that Redis Enterprise’s connection pool manager supported every Redis command and that this would resolve their issue for the short term. In under three weeks, Redis Labs and Netmeds worked together to successfully migrate its caching and session store operations from ElastiCache to Redis Enterprise. Netmeds’ entire application system runs on AWS, with Redis Enterprise now sitting just behind the web app servers. With connection pooling in place, database downtime issues have been eliminated. Long term, Netmeds sees clustering and sharding as the permanent solution for high availability and performance. To that end, Pandit and his team have begun re-architecting their platform to get around its current clustering limitations. They are looking forward to unleashing the power of Redis Enterprise’s clustering and sharding features.

运营影响

Redis Enterprise’s automated management, scaling, and failover capabilities allow Netmeds to provide its e-commerce pharmaceutical platform without interruption.
Since deployment, Netmeds has experienced zero downtime and zero issues.
Redis Labs’ knowledgeable support engineers were able to migrate database operations from ElastiCache to Redis Enterprise and eliminate all downtime issues.

数量效益