-
Hardware failure 2 nights ago resulted in the database being emergency migrated to new hardware. It's a big database so it takes time (on top of the time to detect it).
Then last night a network failure during "emergency scheduled network maintenance":
The attachment shows how it affected the SLO... 1h32m downtime 2 nights ago, and 54m last night for a total of 2h26m downtime. Still within 99.6% uptime this month... and well within 99.95% this year... but yeah, sometimes hardware and networks have issues.
1 Attachment
What happened last night and the night before? I was getting a CloudFlare error message