# Limitations & Data Sizing Guidelines

Valsight allows you to build complex models and the performance of the system is dependent on the actual model and especially the amount of data that is created during multi-dimensional calculations. This page tries to give an overview of tested configurations.

Topic | Supported | Enforced | Notes / Recommendation |
---|---|---|---|

Max Excel File upload size | < 20 MB | 20 MB | Typical Valsight models use aggregated data which fits in ~1 MB files. |

Max number of rows per Sheet in XLS file | 100000 | Excel Maximum = 1048576 rows | Structure your data to not use more than 100.000 rows. Parsing Excel files needs plenty of memory. If working with larger workbooks is mandatory, make sure to run on at least 16 GB. Some frontend parts such as the "Table View" component perform poorly in the browser. Use the server-side table viewer from the admin config interface instead. |

Maximum number of columns per sheet | 26 | Excel Maximum | There is no enforced limit, but the frontend is tested to work with this amount. Note that Excel parsing is dependent on the number of rows and columns. The largest supported configuration contains 100.000 x 26 = 2.6M data cells. |

Level Member | 20000 | not enforced | There is no enforced limit, but the frontend is tested to work with this amount. |

Max number of Dimensions / Hierarchies | 10 | not enforced | There is no enforced limit, but the frontend is tested to work with this amount. |

Number of data rows per Cube | 100000 | not enforced | A cube is the result of a calculation or the underlying data table for a data node. The upper bound of possible values can be calculated as the cross product of all level members, however most real-world examples have very sparse cubes. There is no enforced limit. |

Number of rows per OData/JDBC data source | Dependent on server hardware | not enforced | The default value is set to 100.000. It can be changed in the Valsight settings > Settings under 'System' and the feature flag 'featureFlags.limitODataAndJDBCRowsUpTo'. |

Maximum number of levels per Hierarchy (Hierarchy Depth) | 7 | not enforced | There is no enforced limit, but the frontend is tested to work with this amount. |

Maximum number of data rows per Assumption Line-Item | 1000 | enforced, but configurable | This limit is configurable in the admin section. Higher values may reduce the performance of the frontend. Extending the possible number of rows that can be entered via assumptions is a current focus topic of the Valsight R&D team. |

Number of concurrent users | Highly dependent on the model & user behavior, and CPU and RAM of the server. | You can estimate the amount of concurrent users by estimating the runtime of an average simulation run and the amount of simulations necessary to sustain a certain user behaviour. Models, where the simulation takes two seconds, can therefore roughly sustain 0.5 simulation per CPU core. For example, assuming a user performs a simulation every 10 seconds, an entry level 4 core system can provide enough performance for 0.5*4*10 = 20 Please note that only changes to an assumption count as simulation. | |

Number of Models | 100 | not enforced | The amount of present models has no negative performance impact. |

Number of Workspaces | 200 | not enforced | The amount of simulation workspaces has no negative performance impact. |

Number of Assumptions (Groups) per Workspace | 30 | not enforced | There is no enforced limit, but the frontend is tested to work with this amount. |

Number of Assumption Line-Items per Workspace | 120 | not enforced | There is no enforced limit, but the frontend is tested to work with this amount. |

Number of Template Charts per Model | 20 | not enforced | There is no enforced limit, but the frontend is tested to work with this amount. |

Sequence Dimension | 0-99999 | enforced |