Option in Atlas -
Available for M10+ clusters using General storage class.
- Can specify a minimum cluster size
- Can specify a maximum cluster size
- Can choose to enable/disable downsizing
Works on a rolling basis. Process does not incur any downtime.
Cluster Tier scaling Beta
Metrics taken into account for Cluster Tier scaling up -
If one of below is true
- cpu util ........ >75% for past 1 hour
- memory util .. >75% for past 1 hour
Metrics taken into account for Cluster Tier scaling down -
if both of below are true
- based on data collected in past 72 hours, desired cpu and memory util won't exceed 50%
- cluster has not been scaled down in past 72 hours
Storage Tier scaling
When Disk usage exceeds 90% -
AWS/GCP - increase capacity such that the disk usage becomes 70%
Azure - double the capacity
No comments:
Post a Comment