CloudVPS (openstack) uses a micro services architecture with many scattered components. Finding relevant log entries is often a complex task.

Examples:

• in codfw1dev we have now 3 cloudcontrol servers (API servers, control plane). Each API request can hit several different log files (keystone, neutron, nova)
• in eqiad1 we have 2 web servers, same problem
• we have 2 cloudnet servers per deployment
• etc

Log aggregation or any other facility should be considered to improve our ability to manage these systems.