---
title: "Performance Impacts"
canonical: "https://help.agiloft.com/space/HELP/810090501/Performance+Impacts"
format: markdown
---
If you’re looking to improve your system performance, use this page to guide you in reviewing your configuration. These recommendations can improve responsiveness, reduce load times, and enhance user experience.

Many of the biggest performance impacts come from linked fields, saved search filters, and rules. You might need to do some exploratory investigation to see how these elements are interconnected in your KB as part of your review.

## High Impact Configuration

Saved searches, default values, and linked records can use more resources than expected.

#### Search Conditions

Some search operators are much more resource-intensive than others, particularly if used in saved searches that run frequently as filters or as part of automated actions or rules.

- Avoid broad filters using `contains`, especially on text fields.
- If you need to use `contains`, combine it with the filter `AND doesn’t equal NULL`, especially for text fields. This skips fields that are blank.
- Make sure to use the correct operator when comparing two linked fields with multiple values enabled (MVE):
  - The fastest operator is `overlaps with items in, ~==`.
  - The second best operator is `fully contains, ~=`.
  - **Avoid **using is `fully contained in, <<`. This is significantly slower than `fully contains` and searches less efficiently.
  - **Do not **use `equals to, ==` or `doesn’t equal !=`.
- For Duplicates and First/Last filters, avoid using Date/Time fields, Floating Point fields, and fields that require a unique value.

> ✅ Saved searches have a dedicated Performance tab, which logs each slow run of the search. From the Manage… page of saved searches, yellow or red icons indicate at a glance which saved searches have logged slow runs recently, so you know which searches to investigate and optimize.

#### Default Values in Linked Fields with Multiple Values Enabled

Always avoid setting default values for MVE fields while the page is loading, such as during record creation. Even if you apply a search filter, this process adds significant delay.

- If you need to set default values, consider:
  - Using an action button to populate them from the record or table view, instead.
  - Using a rule to populate them when the record is saved. However, this can still impact performance, especially if this action runs in the foreground.
- Even if you aren’t using AI features, check the settings for the All Internal Entities fields in both the Contract and AI Project tables. If a default value is configured in either field, or is set by a rule or action when records are saved, consider removing these to improve performance.

#### Linked Record Edits Triggering Other Linked Record Edits

There may be cases where you want a modification to a record to automatically trigger modifications to a linked record, or to trigger the automatic creation of new linked records. This process can cause exclusive locks that remain until all the activity is completed.

To avoid this, use a background table to manage the process instead of tying the two directly to each other, so that the initial modification writes data to the background table, and that in turn indirectly triggers the modification or creation of another linked record.

> 📝 This is especially important in environments with many users working concurrently on the same tables.

## Display Options

Simple display options can have a noticeable impact on the time it takes to load records and pages in your KB.

#### Display of Linked Fields

Some display options for linked fields have a significant impact on loading times, and responsiveness when editing records.

- For a linked field with a lot of available values, avoid displaying it as a list of values. Loading the whole list to be available in the drop-down causes slower overall loading for the page, and also becomes resource intensive. This is exacerbated if there are multiple such linked fields for a record. Instead, use the hyperlinked box with lookup.
- When a linked field is used only to display information, use a view-only option like plain text or hyperlinked view only.

#### Layout Design

Avoid using more than two columns on any layout tab. Even a single tab with more columns significantly affects performance for the entire table.

#### Number of Fields

Minimize the total number of fields in each table. Fields can be automatically updated when saving, so delete unused fields.

When saving a linked record, all the fields in the linked set are updated, so make sure that linked sets include only the fields that are necessary. Remove unused fields from linked sets to improve performance.

If you have unused tables, deactivate them so that any fields linking to those tables are also automatically disabled. This reduces background processing load.

## Resource Usage Behind the Scenes

You can improve performance by carefully reviewing the activity that goes on in the background of your workflows, to cut down on duplicative or redundant processes and streamline your automation.

#### Excessive Rule Triggers

Always include filters on rules that are triggered when records are edited or saved, so that the rules only run when there are changes to specific relevant fields in the records. This is especially important for edits that update linked fields or trigger actions involving formulas.

In general, avoid resetting default values or refreshing linked sets unless it’s absolutely necessary. Refreshing linked sets is particularly resource-intensive, and it should only be done if there is a real risk of creating orphaned links.

#### Excessive Saves

Review the History tab for a new record that has been saved, and look for excessive saving. For example, you might have action buttons that save both before and after running their actions, or you might have a series of rules that run and each save their changes. 

Consider removing extra saving from action buttons, and consolidate multiple small rules into a single main edit rule, when possible.

#### History Tracking

When you save a record, a significant portion of the time spent saving is actually spent creating the history entry. Reducing the size and frequency of history entries significantly reduces save times.

- In the History field, limit the overall number of fields tracked. When tracking linked records, track only the ID field of the linked set, rather than tracking every field in the set, unless other fields are used in rule conditions that evaluate which fields were most recently modified.
- Watch for rules that create history entries for all saves. These create a history entry even when no change was actually made to the record. In general, rules should have history tracking set to “Only when at least one field tracked by history is changed,” and both exclusion options enabled unless your rule is specifically intended to modify linked records.
- Don’t track history for Append-Only Text fields; functionally, they track their own history. Likewise don’t track history for fields that store instructional text for the user.

#### Time-Based Rules

Avoid running resource-heavy time-based rules during peak usage hours.

Always apply a saved search to time-based rules if at all possible. Running these rules across all records is extremely resource intensive. In particular, never set a time-based rule to run multiple times a day unless you can apply a saved search to limit its scope. 

If a high-frequency time-based rule often produces no results, run it less frequently or apply a narrower filter.

#### Indexing

Review and maintain indexes regularly, especially after major configuration changes or substantial data imports.

You can request Support assistance to analyze slow queries and propose additional indexes. Focus on the process tables and the most-used background tables, such as Contracts, Attachments, Approvals, Communications, Companies, and Locations.

#### Data Imports and Linked Fields

Avoid importing data directly into linked fields. Instead, import into text fields first, then use an Update action or conversion to populate the target linked fields, so you can do these with a rule run in the background.

Consider the volume of updates as well. Small-scale, single-record imports behave differently than bulk imports of thousands of records.

#### Calculated Fields

Avoid division by zero in Calculated Result fields and other fields performing calculations. Even hidden fields can generate warnings and affect performance.

Where possible, use ternary logic to prevent these errors by performing calculations only if the denominator is not zero. For example: `X != 0 ? Y / X : 0`

#### Text and Short Text Fields

Configure your fields to use an appropriate maximum field length, rather than setting it as high as possible. When text field changes are tracked in history, each change is stored using the full configured size of the field, even if not all the space is used.

Don’t set select a data role of “As HTML” unless you need to use HTML tags in the field. Fields that are stored as HTML take up additional space for each logged history entry. This applies to Text, Short Text, and Append-Only Text field types.

## One-time Actions

#### Deleting Fields

If the field is tracked in History, and there are a lot of history entries for the field, remove the field from History tracking before you delete the field. This makes the deletion much quicker.

When deleting a field from a linked set, never add another field in the same operation. Delete the field first and then go back to add the new field.

#### Updating Changed Fields Only

Set the global variable record_update_changed_fields_only to Yes to reduce system impact when saving records, creating records, and touching records via automation and rules. With this set to Yes, the system updates only the fields that were modified, instead of rewriting the entire row int he database.

#### Related articles

> Macro (contentbylabel)