February 2024 Summaries

Developing a cross-platform chat app with built-in moderation is streamlined using Momento, which supports various client SDKs and minimizes development costs while enhancing user flexibility. The app's core feature is a moderated chat API that filters profane messages and images using the bad-words library and Amazon Rekognition, respectively, and translates text into multiple languages via Amazon Translate. This setup allows users to communicate in different languages by subscribing to language-specific topics, with all client-side applications relying on a centralized backend. The chat app supports a range of platforms, including web, iOS, Android, Flutter, Unity, and React Native, each using different SDKs to manage tasks such as obtaining Momento auth tokens, subscribing to topics, and posting messages. The implementation requires minimal server-side code, primarily encapsulated within a Lambda function, and eliminates the need for WebSockets, thereby reducing dependencies and development efforts across platforms.

Momento's Developer Ecosystem team developed RoboMo, a Discord bot designed to assist users by answering questions related to Momento developer documentation and blog posts through a combination of Momento Vector Index and OpenAI. Although the Momento Vector Index was deprecated in April 2024, the article provides a detailed walkthrough of the bot's creation, which includes setting up a Discord bot application, utilizing the discord.js library to interact with the Discord API, and deploying the bot using AWS Fargate for continuous operation. RoboMo listens for mentions and utilizes a question-answering system to provide responses, with all related code available on GitHub for those interested in creating similar bots.

AWS Lambda, a serverless computing service, enables developers to focus on code execution rather than server management, offering benefits like scalability and cost-efficiency. However, it presents unique challenges, especially regarding long-lived connections and keepalive mechanisms. This blog explores AWS Lambda's execution context reuse, which can preserve global variables across invocations but does not allocate continuous compute resources, impacting tasks like setInterval. At Momento, implementing gRPC keepalive checks in AWS Lambda revealed issues with network communication due to Lambda's dormant state between invocations, leading to timeout errors and the need for reconnection. These findings prompted the decision to disable keepalive pings in Lambda, reducing client-side timeout errors but also delaying the detection of dropped connections. The experience emphasized the importance of adapting to serverless platform behaviors, such as execution context freezing and thawing, and highlighted the necessity for serverless-specific configurations and detailed monitoring to maintain service reliability.

In episode 5 of the Observe It podcast, host Daniela Miao and guest Alex Kehlenbeck, a Distinguished Software Engineer at Lightstep, discuss the complexities and fragmentation of the observability industry, with a particular focus on the challenges posed by high cardinality in metrics. They explore how high cardinality affects the usability and cost of observability tools and discuss the historical difficulties in managing it. The conversation highlights advancements in the OLAP space that could improve observability tools and emphasizes the need for a shift in the industry’s mindset to enhance user experience. Alex discusses pricing model issues, noting customer reluctance to adopt alternatives based on data bytes rather than cardinality, and describes how Lightstep uses OLAP-inspired storage systems to make observability more cost-effective. The episode also covers the potential for disruption in pricing models, the importance of optimizing for cost and performance, and concludes with a discussion on the need for smaller players in the industry to drive change and the future convergence of metrics, traces, and logs for a seamless user experience.