Sizing and scalability guidelines for Gateway Servers
This topic provides the sizing and scalability guidelines for the hardware capacity that is required for Gateway Servers. The following factors affect the sizing drivers:
- Size of your environment
- Number of Agents from which data is collected
- Data retention period
The following table provides the recommendations for the hardware capacity according to the environment size.
Operating system | Size of your environment | Processor cores | RAM (4GB/core) | IOPS | Storage (in GB) for | ||
---|---|---|---|---|---|---|---|
Gateway Server, UDR, and VIS files | UDR files | VIS files | |||||
Windows | Small (up to 1000 servers) | 4 | 16 | 200 | 160 | 150 | 10 |
Medium (up to 3000 servers) | 8 | 32 | 600 | 480 | 450 | 30 | |
Large (up to 5000 servers) | 12 | 48 | 1000 | 790 | 740 | 50 | |
Linux | Small (up to 1000 servers) | 2 | 8 | 200 | 160 | 150 | 10 |
Medium (up to 3000 servers) | 4 | 16 | 600 | 480 | 450 | 30 | |
Large (up to 5000 servers) | 6 | 24 | 1000 | 790 | 740 | 50 |
These calculations are based on the following assumptions:
- Metric resolution - 60 minutes
- Processing window - 4 hours
- UDR data retention - 1 month
- VIS data retention - 3 months
- UDR spill interval - 15 minutes
- UDR data compression enabled - size after 2 days
Data retention
The data collected by Capacity Agents (in the UDR format) is periodically transferred to the Gateway Server where it is automatically processed into hourly intervals and saved in text files called VIS files. BMC recommends the following guidelines for the retention period of UDR and VIS files:
- Keep about three months' worth of VIS files. These files are useful if you need to "recover" data. For more information, see Recovering data.
- For UDR and VIS files, keep three months of data. Out of this data, the oldest two months data can be compressed to save space.