What Sizing Options Are Available to You?
When you create an Oracle Analytics Cloud instance for your production or non-production environment, you either specify the number of Oracle Compute Units (OCPUs) you want to deploy or the number of people you expect to use the service.
- How Many OCPUs Do You Think You’ll Need?
- How Many People Do You Expect to Use the Service?
- What's the Difference Between Production and Non-Production Environments
How Many OCPUs Do You Think You’ll Need?
Oracle Analytics Cloud offers a range of compute sizes (OCPUs) to suit different scenarios and environments. The larger the compute size, the greater the processing power. If you're not sure which size to use, contact your sales team to discuss sizing guidelines.
The compute size you select also determines some configuration limits for the different types of content that users can create:
-
Data visualizations
-
Classic analyses and dashboards
-
Classic pixel-perfect reports
For example, limits such as the maximum number of input rows you can return from a data source query or the maximum number of rows you can download from a report to a file (for example, when you export to a CSV file).
Limits Querying Data (Data Visualization Workbooks, Classic Analyses and Dashboards)
When you query a data source for visualizations or classic analyses and dashboards, the compute size determines the maximum number of rows that are returned from the data source.
Which compute size do you think you’ll need? | Limits when querying data for visualizations, analyses, and dashboards | |
---|---|---|
Max input rows returned from any data source query | Query timeout (seconds) | |
1 OCPU (non-production only) |
125,000 |
660 |
2 OCPU |
2,000,000 |
660 |
4 OCPU |
2,000,000 |
660 |
6 OCPU |
2,000,000 |
660 |
8 OCPU |
2,000,000 |
660 |
10 OCPU |
2,000,000 |
660 |
12 OCPU |
2,000,000 |
660 |
16 OCPU |
4,000,000 |
660 |
24 OCPU |
4,000,000 |
660 |
36 OCPU |
4,000,000 |
660 |
52 OCPU |
4,000,000 |
660 |
Limits Displaying Data (Data Visualization Workbooks, Classic Analyses and Dashboards)
When you display data in visualizations or classic analyses and dashboards, the compute size determines the maximum number of summarized rows returned from the data source that are displayed.
Which compute size do you think you’ll need? | Limits when displaying data in visualizations, analyses, and dashboards |
---|---|
Max summarized rows returned from any data source query | |
1 OCPU (non-production only) |
125,000 |
2 OCPU |
500,000 |
4 OCPU |
500,000 |
6 OCPU |
500,000 |
8 OCPU |
500,000 |
10 OCPU |
500,000 |
12 OCPU |
500,000 |
16 OCPU |
1,000,000 |
24 OCPU |
1,000,000 |
36 OCPU |
1,000,000 |
52 OCPU |
1,000,000 |
Limits Exporting Data (Data Visualization Workbooks)
When you export data from a data visualization workbook, the compute size determines the maximum number of rows you can export, number of parallel exports, and queue size for incoming export requests. If you regularly exceed export limits, you can scale up to a larger compute size or reduce the number of parallel export requests.
There are different row limits for formatted and unformatted data.
-
Unformatted data limit: Comma Separated Values (CSV)
-
Formatted data limit: Microsoft Excel (XLSX)
*The maximum row limits shown in the table are based on exports that contain up to 40 columns. Additional columns will impact the maximum number of rows you can export.
Data exports are expensive operations and have a direct impact on the overall system performance. The impact on system performance increases with the number of rows and columns that you export. Oracle recommends that you export large amounts of data during non-peak hours to reduce any performance impact.
Which compute size do you think you’ll need? | Limits when exporting data from data visualization workbooks | |
---|---|---|
Maximum number of rows exported to CSV* | Maximum number of rows exported to Microsoft Excel (XLSX)* | |
1 OCPU (non-production only) |
125,000 |
25,000 |
2 OCPU |
2,000,000 |
25,000 |
4 OCPU |
2,000,000 |
25,000 |
6 OCPU |
2,000,000 |
25,000 |
8 OCPU |
2,000,000 |
25,000 |
10 OCPU |
2,000,000 |
25,000 |
12 OCPU |
2,000,000 |
25,000 |
16 OCPU |
4,000,000 |
25,000 |
24 OCPU |
4,000,000 |
25,000 |
36 OCPU |
4,000,000 |
25,000 |
52 OCPU |
4,000,000 |
25,000 |
Limits Exporting Data (Classic Analyses and Dashboards)
When you export data from analyses and dashboards, the compute size determines the maximum number of rows you can export, number of parallel exports, and queue size for incoming export requests. If you regularly exceed export limits, you can scale up to a larger compute size or reduce the number of parallel export requests.
There are different limits for formatted reports and unformatted reports.
-
Unformatted report limits: formats such as CSV, Excel, XML, and Tab Delimited.
-
Formatted report limits: formats such as PDF, Excel, Powerpoint, and Web Archive/HTML.
When two pivot views are laid out side by side in a union the formatted export limit will be 20,000 rows.
*The maximum row limits shown in the table are based on exports that contain up to 40 columns. Additional columns will impact the maximum number of rows you can export.
Data exports are expensive operations and have a direct impact on the overall system performance. The impact on system performance increases with the number of rows and columns that you export and the output format. Oracle recommends that you export large amounts of data during non-peak hours or export unformatted data to reduce any performance impact.
Which compute size do you think you’ll need? | Limits when exporting data from analyses and dashboards | |
---|---|---|
Maximum number of rows exported to unformatted reports* | Maximum number of rows exported to formatted reports* | |
1 OCPU (non-production only) |
125,000 |
1,000 |
2 OCPU |
2,000,000 |
200,000 |
4 OCPU |
2,000,000 |
200,000 |
6 OCPU |
2,000,000 |
200,000 |
8 OCPU |
2,000,000 |
200,000 |
10 OCPU |
2,000,000 |
200,000 |
12 OCPU |
2,000,000 |
200,000 |
16 OCPU |
4,000,000 |
400,000 |
24 OCPU |
4,000,000 |
400,000 |
36 OCPU |
4,000,000 |
400,000 |
52 OCPU |
4,000,000 |
400,000 |
Limits Delivering by Email (Classic Analyses and Dashboards)
When you send analyses and dashboards by email, the compute size determines the maximum number of rows you can deliver in a single email. There are different limits for delivering formatted reports and unformatted reports.
-
Unformatted report limits: formats such as CSV, XML, and Tab Delimited.
-
Formatted report limits: formats such as PDF, Excel, Powerpoint, and Web Archive/HTML.
Content delivery by email is an expensive operation and has a direct impact on the overall system performance. The impact on system performance increases with the number of recipients, the number of rows and columns that you send, and the delivery format. Oracle recommends that you schedule deliveries during non-peak hours or change the delivery format to reduce any performance impact.
Which compute size do you think you’ll need? | Limits when delivering analyses and dashboards by email | |
---|---|---|
Max rows in unformatted reports delivered by email | Max rows in formatted reports delivered by email | |
1 OCPU (non-production only) |
2,000 |
1,000 |
2 OCPU |
200,000 |
50,000 |
4 OCPU |
200,000 |
50,000 |
6 OCPU |
200,000 |
50,000 |
8 OCPU |
200,000 |
50,000 |
10 OCPU |
200,000 |
50,000 |
12 OCPU |
200,000 |
50,000 |
16 OCPU |
300,000 |
100,000 |
24 OCPU |
300,000 |
100,000 |
36 OCPU |
300,000 |
100,000 |
52 OCPU |
300,000 |
100,000 |
Data Size Limits (Classic Pixel-Perfect Reports)
The compute size determines several limits associated with generating reports.
Which compute size do you think you’ll need? | Report data size limits when generating pixel-perfect reports | |||
---|---|---|---|---|
Max data size for online reports | Max data size for offline (scheduled) reports | Max data size for bursting reports | Max data size for data generation | |
1 OCPU (non-production only) |
200MB |
500MB |
2GB |
500MB |
2 OCPU |
500MB |
2GB |
4GB |
2GB |
4 OCPU |
500MB |
2GB |
4GB |
2GB |
6 OCPU |
500MB |
2GB |
4GB |
2GB |
8 OCPU |
500MB |
2GB |
4GB |
2GB |
10 OCPU |
500MB |
2GB |
4GB |
2GB |
12 OCPU |
500MB |
2GB |
4GB |
2GB |
16 OCPU |
500MB |
4GB |
8GB |
4GB |
24 OCPU |
500MB |
4GB |
8GB |
4GB |
36 OCPU |
500MB |
4GB |
8GB |
4GB |
52 OCPU |
500MB |
4GB |
8GB |
4GB |
Processing Limits (Classic Pixel-Perfect Reports)
The compute size determines several limits associated with processing reports.
Which compute size do you think you’ll need? | Data model and report processing limits when generating pixel-perfect reports | ||||
---|---|---|---|---|---|
SQL Query timeout for scheduled reports (seconds) | Max rows for CSV output | Max number of in-memory rows in XPT layout | Max number of concurrent scheduled jobs | Max number of concurrent online reports | |
1 OCPU (non-production only) |
1,800 |
1,000,000 |
100,000 |
1 |
2 |
2 OCPU |
1,800 |
4,000,000 |
200,000 |
4 |
16 |
4 OCPU |
1,800 |
4,000,000 |
200,000 |
4 |
32 |
6 OCPU |
1,800 |
4,000,000 |
200,000 |
4 |
48 |
8 OCPU |
1,800 |
4,000,000 |
200,000 |
4 |
64 |
10 OCPU |
1,800 |
4,000,000 |
200,000 |
4 |
80 |
12 OCPU |
1,800 |
4,000,000 |
200,000 |
4 |
96 |
16 OCPU |
3,600 |
6,000,000 |
300,000 |
10 |
320 |
24 OCPU |
3,600 |
6,000,000 |
300,000 |
10 |
480 |
36 OCPU |
3,600 |
6,000,000 |
300,000 |
10 |
720 |
52 OCPU |
3,600 |
6,000,000 |
300,000 |
10 |
1040 |
How Many People Do You Expect to Use the Service?
With Oracle Analytics Cloud, you can opt to specify how many people you expect to use the service. Typically, services have between 10 and 3000 users. Configuration limits for user-based subscriptions are equivalent to those shown here.
Configuration Limits for User-based Subscriptions: Data Visualization Workbooks, Classic Analyses and Dashboards
Limit Description | Limit Value |
---|---|
Limits when querying data for visualizations, analyses, and dashboards | |
Max input rows returned from any data source query | 2,000,000 rows |
Query timeout (seconds) | 660 seconds |
Limits when displaying data in visualizations, analyses, and dashboards | |
Max summarized rows returned from any data source query | 500,000 rows |
Limits when exporting data from data visualization workbooks | |
Maximum number of rows exported to CSV | 2,000,000 rows |
Maximum number of rows exported to Microsoft Excel (XLSX) | 25,000 rows |
Limits when exporting data from analyses and dashboards | |
Max rows exported to unformatted reports | 2,000,000 rows |
Max rows exported to formatted reports | 200,000 rows |
Limits when delivering analyses and dashboards by email | |
Max rows in unformatted reports delivered by email | 200,000 rows |
Max rows in formatted reports delivered by email | 50,000 rows |
Configuration Limits for User-based Subscriptions: Classic Pixel-Perfect Reports
Limit Description | Limit Value |
---|---|
Report data size limits when generating pixel-perfect reports | |
Max data size for online reports | 500MB |
Max data size for offline (scheduled) reports | 2GB |
Max data size for bursting reports | 4GB |
Max data size for data generation | 2GB |
Data model and report processing limits when generating pixel-perfect reports | |
SQL Query timeout for scheduled reports (seconds) | 1,800 seconds |
Max rows for CSV output | 4,000,000 rows |
Max number of in-memory rows in XPT layout | 200,000 rows |
Max number of concurrent scheduled jobs | 4 jobs |
Max number of concurrent online reports | 32 reports |
What's the Difference Between Production and Non-Production Environments
-
Non-production environment: Oracle enables you to deploy a non-production environment with 1 OCPU. A non-production environment is specifically sized and designed for test, development and training purposes. Non-production services aren't intended for daily use, multiple concurrent users, or complex business scenarios.
If you decide you want to keep the content that you create during testing, you can save it to a snapshot and copy it to a production service (minimum 2 OCPU or 10 users). See Migrate Oracle Analytics Cloud Using Snapshots. Alternatively, you can scale up your 1 OCPU environment to between 2 and 8 OCPUs.
-
Production environment: A production environment is designed for daily commercial use. You can scale some production environments up and down. For example, you can scale between 2 and 8 OCPUs and 10 and 12 OCPUs. You can also scale between various user ranges such as 10 - 400 users and 401 - 601 users. See About Scaling.