Difference between revisions of "Fusion Data Store memory requirements"
Jump to navigation
Jump to search
| Line 5: | Line 5: | ||
* The balance between series and observations will also affect memory requirements with datasets where each series has few observations needing additional memory. | * The balance between series and observations will also affect memory requirements with datasets where each series has few observations needing additional memory. | ||
* The figures quoted above are peak memory requirements when updating or replacing existing datasets. Quiescent memory requirements are lower. | * The figures quoted above are peak memory requirements when updating or replacing existing datasets. Quiescent memory requirements are lower. | ||
| + | * Services such as Fusion Registry have additional memory requirements which need to be taken into account when sizing deployment platforms such as the volume of stuctural metadata. | ||
The following video demonstrates some live tests. | The following video demonstrates some live tests. | ||
https://www.youtube.com/watch?v=4nLZNKTN1BE | https://www.youtube.com/watch?v=4nLZNKTN1BE | ||
Revision as of 11:07, 26 January 2022
Fusion Data Store requires a peak of approximately 6GB of server memory per 50 million observations:
120MB per million observations.
Notes:
- The balance between series and observations will also affect memory requirements with datasets where each series has few observations needing additional memory.
- The figures quoted above are peak memory requirements when updating or replacing existing datasets. Quiescent memory requirements are lower.
- Services such as Fusion Registry have additional memory requirements which need to be taken into account when sizing deployment platforms such as the volume of stuctural metadata.
The following video demonstrates some live tests. https://www.youtube.com/watch?v=4nLZNKTN1BE
