Resolve intermittent downtime on production server

[faatihi]

Problem Description

Action plan: separate Database from Application using different server instances. Steps:

• [x] write a script to automate setting up replication in a sandbox
• [x] apply replication script on the test server against training server
• [x] separate application from database into different server instances for the new DR
• [x] find a suitable tool for automatic failover
• [ ] apply replication script on the DR against production server
• [ ] switch production traffic to DR
• [ ] observe and measure effects, improvements, stability of the system for a week
• [ ] if system is stable, even if there's no noticeable improvement, split app and db for production
• [ ] switch traffic back to production

Proposed Solution

Possible Alternatives

Additional Information

[faatihi]

This issue will be closed for two reasons:

1. It's been about 6 months, we haven't had issues regarding downtime. This is attributed to the version upgrades we deployed. The upgrades included a number of performance improvements.

2. We are not moving forward with the splitting of the app and database. We got benchmarks that show there isn't much difference in placing the app and database on separate servers. Also, it could face more performance issues because we will not be able to easily adjust allocation of resources when one server needs more resources than the other in a particular period. Having both on the same server allows for flexibility in tweaking resource allocation as per the system's needs.