High CPU Usage Causing IDE Services Hosted by AWS EC2 Instance Crashes

Problem

Repeated system crashes due to high CPU utilization. Investigation points to the salt-minion service as a primary contributor to these spikes.

Cause

The salt-minion service, part of the Salt stack infrastructure automation platform, appears to be consuming excessive system resources periodically. This issue is compounded by the hosting environment running on an AWS EC2 t3a.medium instance. A single t3a.medium instance provides 2 vCPU / 4 GB in total, provides insufficient resources when multiple services (Salt-minion, IDE Services application, DB, OS processes) are run concurrently. On a burstable instance type, the minimum tier requires the IDE Services application server at 2 CPU / 3 GB RAM plus a separate PostgreSQL at 2 CPU / 2 GB RAM, according to the system requirements.

Resolution

1. Application Resource Confirmation:

   • The salt-minion service is unrelated to JetBrains software.
   • Please confirm that the IDE Services application remains stable and does not contribute to the high CPU spikes and PostgreSQL is running on a separate RDS instance.

2. System Requirements Review: The t3a.medium instance is only sufficient for running the application itself, it lacks headroom for additional services or processes. Please confirm that PostgreSQL running on a separate RDS instance, db.t3.large should be adequate for the database needs.

3. Recommendation: Please consider upgrading the instance type for the host running IDE Services, Salt-minion and other OS-level agents/processes to a larger size (e.g., AWS EC2 t3a.large) to accommodate additional resource requirements.

If the issue persists, please create a Zendesk issue and share the specifications of your IDE Services Server and the Server Logs covering the period when the CPU utilization spiked.