Known system limitations
ХТRF platform, like any other software, has some technical limitations concerning the database size, maximum number of characters, etc. Our QA team regularly tests the system capabilities using the available data and known scenarios on the databases we have access to. However, we can't guarantee that our team can test and confirm all possible scenarios and configurations for a particularly large number of entities.
If you intend to work with large amounts of data, especially if you already have data amount-related problems with other software, please reach out to your main point of contact before proceeding.
It is always better to start on a smaller scale and follow up when the performance is satisfactory.
In particular, please be aware of the following restrictions:
There is a hard limit on uploaded file size - 1GB. Please be aware that some file types, especially compressed files, may have much lower size limits due to complex processing and analysis, most commonly related to CAT tool integrations. It is not possible to estimate this limit beforehand. Excessive processing of such files may degrade system performance and lead to a crash.
There is a limit to importing entities from previous systems - 10,000 entities of each kind, for example, 10,000 previous customers. Please take extreme caution to only import necessary and legally allowed personal data into the system.
There is a soft limit to the number of concurrently logged-in users (between 50 and 100), which is heavily dependent on usage patterns and server performance. Scaling above 100 concurrent users will likely require the multinode setup with horizontal user distribution. Exceeding a certain user amount will degrade application performance.
Some text fields in the system allow you to use formatting and paste text from other sources. The contents of those fields should be limited to 10,000 characters (bytes).
A soft limit of contact persons per single company entry is 100. While the system allows for more, it may degrade system responsiveness.
The size of a single translation project, measured in the number of jobs and activities, is limited. Classic Projects should have at most 50 tasks, 50 activities, 50 language combinations, and an overall storage size of 10GB.
While the system allows you to add custom data fields to certain entities, their amount is limited to 10 for each type. You should not store large amounts of data in the custom fields, as this impacts performance significantly.
XTRF also allows you to add scripted virtual columns, similar to custom fields, but dynamically calculated during view rendering. Using too many or too complex virtual columns will increase the CPU usage of the application and reduce its performance.
XTRF macros can process up to 500 entities.
When using CAT tool integrations, the XTRF responsiveness heavily depends on the third-party system responsiveness. Make sure to reduce the network latency between systems and ensure sufficient capacity for the external system.
In load-heavy Smart Projects (e.g., those where a CAT tool integration, job splitting, and process with more than six steps are enabled simultaneously), we recommend limiting the number of target languages to around 50 for better responsiveness.
While API responses, especially browse, allow for a high number of results at one request, it is better to use pagination and scroll through multiple smaller result sets. Especially with virtual columns, large result sets will be slow to calculate.
The vendor portal has shown poor performance when a vendor has over 100 invoices.