987 points by serverlesssam 6 months ago flag hide 10 comments
john_doe 6 months ago next
Nice post! I've been thinking of implementing serverless architecture in my project as well. Can you share more about the specific services you used to ensure zero downtime?
author 6 months ago next
Sure, I used AWS Lambda for the serverless architecture, supported by AWS API Gateway and DynamoDB for low-latency data access. AWS Lambda helps in automatic scaling, thus ensuring zero downtime. I've also set up CloudWatch events to trigger regular health checks.
author 6 months ago next
In addition to CloudWatch for AWS, you can utilize third-party tools like Datadog or New Relic for observability. Also, Azure Monitor and Application Insights can help you monitor your current implementation using Azure Functions and Cosmos DB.
jane_doe 6 months ago prev next
I am using Azure Functions and Cosmos DB. Any suggestions on keeping a tab on scaling and observability to maintain the zero downtime?
nerd_genius 6 months ago prev next
What role did you find AWS S3 play in this architecture?
author 6 months ago next
S3 was used mainly for storing static files and data that are infrequently accessed. It can be configured for automatic backups and versioning in case any issues arise.
sam_tech 6 months ago prev next
What made you decide on a serverless architecture? Can you share some benefits and concerns for serverless?
author 6 months ago next
Serverless helped us to focus more on app features while saving costs on infrastructure management. The benefits include automatic scaling, cost savings, pay-per-use, and short time-to-market. Concerns include limited control over resources, vendor lock-in, and potential cold starts.
codewalls 6 months ago prev next
Good write-up! Could you share more details on your CloudWatch configuration for health checks?
author 6 months ago next
CloudWatch Events can be configured to run custom health check scripts on a regular basis to monitor logs, application performance, or service uptime, and trigger alarms in case of failures or thresholds reach.