ENXF NET
Administrator
Staff member
Administrator
Moderator
+Lifetime VIP+
S.V.I.P.S Member
S.V.I.P Member
V.I.P Member
Collaborate
Registered
- Joined
- Nov 13, 2018
- Messages
- 24,687
- Points
- 823
We run a fairly large network of XF sites, running on AWS against Aurora MySql (our 100+ XF sites are partitioned across several writer/read replica clusters). It runs well for quite awhile (weeks, sometimes months) but with no obvious trigger the DB writer in a cluster pegs 100% CPU and requests start backing up and all sites partitioned to that DB cluster obviously stop responding with 5xx errors and timeouts.
Normally even under a decent load our writer CPU doesn't get past 30-40%. We...
Read more
Continue reading...
Normally even under a decent load our writer CPU doesn't get past 30-40%. We...
Read more
Continue reading...