In the last day me and @samuele.masetto have done some upgrades and tests both on production and dev environment.
Lessema is deadly slow.
Every bash command or webpage refresh takes tens of seconds or, some times, minutes.
Any kind of improvement or evolutive work is impossible.
There is no evidences of any overload in the VM, no background activities, neither high load average.
My opinion is that it is an infrastructure issues.
Using shared resources (and shared CPUs) contabo can not guarantee any minimal performance.
I remember you all that last summer the whole contabo infrastructure has been unreachable for at least 18 hours, with no official communication in the first 12 hours.
I suggest to
- open a ticket on contabo
- evaluate the migration to a better place as soon as possible