Determine Database Sizing

Configuring and Using Diagnostic Tools › How to Clean Up Data with the CA SOI Toolbox › Determine Database Sizing

Determine Database Sizing

Maintaining the SA Store database is required to sustain consistent performance and maintain product function. Use the data provided in this section that is measured against the disk space and performance capabilities of your database server. You can then determine how often to perform maintenance.

The following tables display estimates of how you can expect the database to grow over time. They project growth over a three-year period for several implementation sizes.

Note: All tables assume that you have defined an SLA for every managed service.

Small (100 services, 1000 alerts per day)

Item	Amount	Average Row Size (KB)	Number of Rows	Disk Space Requirement (MB)
Managed CIs	20,000	.85	60,000	51
Staged CIs	100,000	1.345	300,000	403.5
Alerts	1,095,000	.6	3,285,000	1,971
Alert History		1.062	2,190,000	2,325.78
SLA	100	.3	109,500	32.85
History tables		.65	5,256,000	3,416
Total table size				8,201
Database log file				4,920
Total size				13,121

Mid-sized (500 services, 5000 alerts per day)

Item	Amount	Average Row Size (KB)	Number of Rows	Disk Space Requirement (MB)
Managed CIs	50,000	.85	150,000	127.5
Staged CIs	500,000	1.345	1,500,000	2,018
Alerts	5,475,000	.6	16,425,000	9,855
Alert History		1.062	10,950,000	11,629
SLA	500	.3	547,500	164.25
History tables		.65	26,280,000	17,082
Total table size				40,875
Database log file				24,525
Total size				65,400

Large (1,000 services, 10,000 alerts per day)

Item	Amount	Average Row Size (KB)	Number of Rows	Disk Space Requirement (MB)
Managed CIs	100,000	.85	300,000	255
Staged CIs	1,000,000	1.345	3,000,000	4,035
Alerts	10,950,000	.6	32,850,000	19,710
Alert History		1.062	21,900,000	23,258
SLA	1,000	.3	1,095,000	328.5
History tables		.65	52,560,000	34,164
Total table size				81,750
Database log file				49,050
Total size				130,800

Review the following considerations when interpreting these estimates:

The Amount column lists the actual amount of each entity in CA SOI at the end of the three-year period. The Number of Rows column is larger than the actual amount in most cases.
The SA Store tables not represented in this table are small enough that they have a negligible impact on database size.
The History table row calculation is based on the fact that the three history tables (DBAvailHistory, DBQualityHistory, and DBRiskHistory) create a row every thirty minutes per service. Therefore, the history tables generate 48 rows per day per service.
Plug your numbers into the table (total number of managed and staged CIs, average alerts received per day, and number of SLAs defined) that most closely approximates your implementation size to calculate the projected database growth rate for your implementation.
The database log file is approximately 60 percent of the total database size.

Based on this data, you can make the following assumptions:

At the end of a three-year period, the total disk space consumed by the SA Store database would be between 10,000-100,000 MB or 10-100 GB depending on the number of managed services, CIs, and alerts.
You can expect a growth rate of roughly 2.7 GB per 1000 services per month, or 1.35 GB per 500 services per month, or 270 MB per 100 services per month.
Consider these estimates when planning the frequency of maintenance to help ensure that the SA Store does not grow to an unmanageable size. As a best practice, the row count for all tables should be limited so that SQL queries on any table can complete within a few seconds. For example, if you find that this limit is around 10,000,000 rows per table on your database server, you would have to at least archive and purge the history tables periodically depending on implementation size to keep them from growing beyond this threshold.