Logs, Monitoring and Performance Alerts at CrowdHandler

CrowdHandler utilizes logging and monitoring across our infrastructure and application to gain visibility into operational health, ensure uptime and detect potential issues early. Our logging and monitoring stack includes:

Logging

We use Amazon CloudWatch Logs to capture logs from:

  • AWS Lambda functions - Logs all executions of our Lambda functions that run our core application code. Retained for 12 months.
  • API Gateway access logs - Logs all API calls from anonymous end users and registered client users/administrators. Retained for 12 months.
  • AWS RDS database logs - Logs read/write activity and performance metrics for our MySQL databases. Retained according to RDS log retention settings.
  • AWS CloudTrail logs - Logs all AWS API activity for our account. Retained for 12 months.

These logs allow us to trace usage patterns, troubleshoot software errors or performance issues and monitor for security anomalies. However, we do not have access to logs containing sensitive customer data, only our own infrastructure and application logs.

Monitoring

We use Amazon CloudWatch for monitoring key performance metrics across our AWS resources including:

  • Lambda - Monitoring execution times, error rates and throttling for our serverless functions.
  • API Gateway - Monitoring 5XX/4XX error rates, latency, request counts and throttling for our APIs.
  • RDS MySQL - Monitoring CPU, memory, I/O, swap usage and replication lag for our databases.
  • S3 - Monitoring storage usage and request metrics for our file storage.
  • Route53 - Monitoring DNS query rates and latencies.
  • CloudFront - Monitoring request/data transfer rates, 4XX/5XX error rates and latencies for content delivery.

Performance Alerting

We have set up CloudWatch alarms on critical metrics so we receive notifications when thresholds are crossed for:

  • Elevated application error rates (Lambda, API Gateway)
  • Database CPU/memory usage over 75%
  • Replication lag over 2 minutes
  • CloudFront 4XX/5XX error rates over 2%
  • Lambda function duration over 3 seconds
  • S3 Burst Balance in excess of 25% for over 30 minutes

These alerts allow us to respond quickly in case of software errors, infrastructure issues or traffic spikes that could potentially impact our service. On-call team members are paged through an alert notification system 24/7/365.

CrowdHandler's observability stack providing logging, monitoring and alerting helps ensure continuous insight into the health, performance and security of our infrastructure and applications. While we do not log or monitor customer data, we are able to gain visibility into key metrics across the systems supporting our service so we can minimize disruptions and resolve any issues rapidly. Our alerting system guarantees a quick response in the event of software or infrastructure problems at any time.

Please let us know if you have any questions about our logging, monitoring and performance alerting policies or require any further details.