Loading
Salesforce now sends email only from verified domains. Read More
About Salesforce Data 360
Table of Contents
Filter by (0)Add
Select Filters

          No results
          No results
          Here are some search tips

          Check the spelling of your keywords.
          Use more general search terms.
          Select fewer filters to broaden your search.

            Search all of Salesforce Help
            Search all of Salesforce Help
            . Customer Data Platform Limits and Guidelines

            You are here:

            1. Salesforce Help
            2. Docs
            3. Data 360

            Customer Data Platform Limits and Guidelines

            Guidelines define best practice recommendations to optimize the adoption of Data 360 by Customer Data Platform customers for best performance. Limits are boundaries beyond which features are unavailable, performance is throttled, or usage billing charges are applied. This article is about the legacy Customer Data Platform license, which is no longer available for purchase.

            Some limits and guidelines vary by edition or org type, such as when using Data 360 in a Developer org.

            Note
            Note Some limits, guidelines, and usage billing calculations differ depending on your org’s license.

            For billing information, refer to this additional document if your Data 360 org is operating under a Customer Data Platform license.

            For a list of documents related to other licenses, see this document instead.

            Contact your account executive if you aren’t sure which documents apply to you.

            General Guidelines and Limits

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            Location of data storage    
            Number of anonymous profiles
            • Total Number of Known Profiles x 5
            		   For example, if you have 100 known profiles in your account, you can have 500 anonymous profiles.
            Number of Data 360 permission set licenses
            • 500
            		 Yes Learn about Data 360 Standard Permission Sets, Permission Set Considerations, and Salesforce Features and Edition Allocations.
            Number of data spaces
            • Depends on the number of licenses purchased
            • When using Data 360 in a Developer org: 1
            		  

            Learn more about licenses in Data 360 Standard Editions and Licenses.

            In developer orgs, the default data space is the only data space.
            Data Spaces permission sets
            • 1 per data space
            		 Yes

            Data spaces permission sets can’t be included in permission set groups.

            Data spaces permissions sets aren’t counted against the Number of Data 360 Permission Set Licenses limit.

            Unsupported features
            • Data Kits: Marketing Cloud Engagement data streams aren’t installable
            		 Yes  
            Unsupported orgs
            • Professional Edition orgs
            • Government Cloud
            • Sandbox orgs
            		 Yes  
            Data Cloud One connections Not available Yes  

            Activation Guidelines and Limits

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            Scheduled activation publish frequency
            • Minimum: 12 hours
            • Maximum: 24 hours
            		   Learn more in Review Publish History of Your Activation.
            Total number of attributes
            • 100
            		   Learn more about attributes in Create Activation for a Segment.
            Number of segments to an activation target
            • 1
            		 Yes Each segment can be sent to an activation target one time. Learn more in Create Activation for a Segment.
            Number of activation targets total
            • 300
            • When using Data 360 in a Developer org: 15
            		   An activation target is information stored for a given activation platform, which is the location that a segment’s data is being sent to during the activation process.
            Total number of activations
            • Total Number of Contracted Segment Publishes x 4
            		   For example, if you purchased 100 segments, you can create up to 400 activations.
            Total number of activations with related attributes
            • 10,000
            		 Yes Use up to 100 activations with related attributes. After reaching that limit, choose to delete related attributes from existing activations or create activations with only direct attributes.
            Total number of related attributes per activation
            • 300 across no more than 4 data model objects in the same path
            		 Yes Learn more in Considerations for Selecting Related Attributes in Data 360 Activations.
            Maximum records in a segment with related attributes
            • 150 million
            		 Yes If the segment has more than 10 million records and contains related attributes when you publish, the segment doesn’t activate. Learn more in Considerations for Selecting Related Attributes in Data 360 Activations.
            Maximum levels away from the Activation Membership DMO with related attributes
            • 5
            		 Yes Choose attributes from DMOs 5 levels away from your selected Activation Membership DMO. Learn more in Considerations for Selecting Related Attributes in Data 360 Activations.
            Maximum number of related attribute records selected
            • 25
            		 Yes Select the number of records sent for a given attribute in your Activation payload.
            Number of ad audience activations to ad audience partners
            • 3 ad audience activations
            		 Yes

            Additional add audience activations are available for purchase.

            The ad audience partners are Google Ads, Meta (Facebook), and Amazon Ads.

            One ad audience allows a customer to activate one segment to an ad audience partner.

            For example, sending one segment to a Meta account and one segment to a Google account equals two ad audiences.

            If a customer purchases 10 additional ad audience activations, then the total number will be 13 since it comes with 3 and 10 were purchased.

            Number of activation records processed per org per year for each ad audience
            • 85 billion
            		 Yes For example, customers with three ad audiences can consume up to 255 billion activation records processed in a 12-month period.
            Maximum number of characters for file storage activation
            • 200
            		 Yes  

            Calculated Insights Guidelines and Limits

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            How often calculated insights are processed
            • Multiple times a day
            		  

            Insights are processed one time a day to multiple times a day depending on the volume of data and complexity of queries.

            Calculated insights don't run if there are no changes to source data, object mappings, or configurations, or if you’ve already reached the per-insight oo per-org process limit.
            Maximum number of times a calculated insight can be manually processed in any 24-hour period
            • 30 per calculated insight
            		 Yes  
            Maximum calculated insights execution time
            • 2 hours
            		  

            Insights that run for longer than 2 hours may be terminated by the system.

            If a calculated insight runs for longer than two hours, review your data shape and optimize it.

            Maximum number of nested calculated insights
            • 4
            		 Yes Refers to how many existing insights you can use within a new insight.
            Maximum number of dimensions per calculated insight
            • 10
            		 Yes A dimension contains a qualitative value, such as a product name, date, or profile ID.
            Maximum number of measures per calculated insight
            • 50
            		 Yes A measure is a quantitative aggregated value, such as an average or total amount.
            Total number of calculated insights per tenant
            • 300
            		 Yes A tenant is a specific org or instance in Salesforce.
            Maximum number of real-time insights
            • Real-time insights aren't available
            		 Yes  
            Maximum number of real-time insight fields (dimensions and measures)
            • Real-time insights aren't available
            		 Yes  

            Data Actions Guidelines and Limits

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            Total number of data actions per tenant
            • 10
            		 Yes  
            Total number of data action targets per tenant
            • 3
            		 Yes This limit is enforced by the UI.
            Total number of rules per data action
            • 10
            		 Yes Refers to the number of filter rules in a data action definition.

            Data Explorer Guidelines and Limits

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            Total number of columns queried at a time
            • 10
            		 Yes  
            Total number of rows displayed per page
            • 100
            		 Yes  

            Data Federation Guidelines and Limits

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            Number of external data lake objects allowed per org
            • 10
            		   The profiles in the external tables account towards the billable profile count limits and any extra profiles beyond the contracted entitlements would incur overages.

            Data Graph Guidelines and Limits

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            Maximum number of data graphs per org
            • 2
            		 Yes  
            Maximum number of draft data graphs per org
            • 5
            		    
            Scheduled data graph refreshes per day
            • 1
            		 Yes You can schedule data graphs to refresh every four hours, daily, weekly, or monthly. You can also refresh them manually.
            Maximum number of records across all data model objects and calculated insights included in a data graph
            • 200 million
            		   Including objects with more than 200 million records total in a data graph impacts data graph refresh times and query response times.
            Maximum size of a real-time data graph
            • Real-time data graphs aren't available
            		 Yes  
            Maximum number of objects per data graph
            • 25
            		 Yes The data graph’s primary data model object, related data model objects, and calculated insights count toward this limit.
            Maximum number of fields per data model object that can included in a data graph
            • 50
            		 Yes Primary key, foreign key, and key qualifier fields aren’t included in this limit.
            Maximum number of engagement events in a real-time data graph
            • Real-time data graphs aren't available
            		 Yes  
            Maximum age of engagement data included in a real-time data graph
            • Real-time data graphs aren't available
            		 Yes  
            Maximum number of measures per calculated insight that can be included in a data graph
            • 5
            		 Yes  
            Maximum number of total fields and measures that can be included in a data graph
            • 200
            		 Yes Primary key, foreign key, and key qualifier fields aren’t included in this limit.
            Maximum number of levels that data model objects (DMO) can be nested below the primary DMO in a data graph
            • 5
            		 Yes Levels are represented as a tree structure in the data graphs editor. The nesting tree including the primary DMO can have a total of 6 levels.
            Maximum number of records allowed by the recency filter
            • 100
            		 Yes  
            Maximum amount of time available for the time filter
            • 720 hours (30 days)
            		 Yes  

            Data Ingestion Guidelines and Limits

            General

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            Maximum number of fields per data stream
            • 1,050
            		 Yes  
            Maximum number of single field types in a data stream
            • 800
            		 Yes  
            Total data size limit
            • When using Data 360 in a Developer org: 100 GB
            		    
            Total number of CRM custom objects
            • 20
            		   This limit applies to the number of custom objects ingested from each connected CRM org.
            Total Number of CRM orgs that can be connected to Data 360
            • 5
            		    
            Total number of data models
            • 7,500
            • When using Data 360 in a Developer org: 300
            		 Yes  
            Total number of data streams
            • 5,000
            • When using Data 360 in a Developer org: 100
            		 Yes  
            Total number of data lake objects (DLOs)
            • 5,000
            • When using Data 360 in a Developer org: 100
            		 Yes This limit applies to data lake objects created as part of data streams or in a standalone manner.
            Total Number of Marketing Cloud Personalization (Interaction Studio) datasets per instance
            • 5
            		    
            Maximum number of rows for a Marketing Cloud Engagement Data Extension full extract
            • 100 million
            		 Yes  

            GCS, Azure, and S3 Bulk Ingestion

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            Optimal number of file rows when creating S3 data streams
            • 1,000
            		   When creating an S3 data stream, you may need to use a file smaller than the data you’d like to ingest. The smaller file should contain the same headers and about 1,000 rows. Once the data stream is created, larger files can be placed within the file path, maintaining that they fall within set limits.
            Optimal size per CSV file
            • 500 MB
            		   Sometimes, particularly for Day 0 loads, it’s necessary to move large data volumes. Recommended limits are: 200 GB max per file and 1,000 files max per scheduled run. These two parameters aren’t intended to be stretched to their respective maximums concurrently. Rather consider some tradeoff between the two extremes and refer to the maximum data size guidelines for more context.
            Optimal size per parquet file
            • 50 GB
            		  
            Maximum number of files per scheduled run
            • 1,000
            		  
            Maximum size per uncompressed CSV file
            • 200 GB
            		    
            Maximum size per compressed CSV file
            • 250 MB
            		   Compressed CSV files aren’t splittable, meaning that they don't allow for parallel processing and their usage is therefore discouraged. If sending in a compressed file, it must be a single file per zip.
            Maximum size per uncompressed parquet file
            • 200 GB
            		   We recommend you to use the parquet compression capabilities to reduce file size.
            Maximum size per compressed parquet file
            • 100 GB
            		   Compressed parquet files are splittable and support parallel processing of ingestion. There are several native compression options. Learn more about supported compression formats in Supported File Formats in Data 360.
            Maximum data size for upsert
            • 50 GB
            		  

            When a data stream is on upsert, it’s best not to exceed a 50-GB aggregate across all files in the single run. This number is strongly influenced by the number of distinct data dates present in the table, the overall size of the table, and several other factors such as the number of columns. As these variable inputs increase you may experience performance degradation.

            For larger target tables and larger upsert ingestions, we strongly recommend using engagement tables and selecting an engagement date field that contains many distinct dates. Not using engagement nor selecting a high cardinality date field may cause performance degradation.
            Maximum data size for full refresh
            • 1,000 GB
            		   Data size is the sum of all file sizes for a single datastream job. When a data stream is set for full refresh, it’s best not to exceed 1000 GB for all files in a single run. This number is strongly influenced by the number of distinct data dates present in the table, the overall size of the table, and several other factors such as the number of columns. As these input variables increase you may experience performance degradation.
            Maximum number of characters for a single cell 70,000 Yes  

            Local File Upload

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            Maximum number of characters for a single cell in a local file upload
            • 70,000
            		 Yes  
            Supported delimiters
            • We currently only support the comma (,) delimiter.
            		 Yes  
            Header row requirements
            • CSV files require a header row name for every field. This requirement is the same format as what's used for Cloud Drive uploads.
            		 Yes  
            Unsupported refresh of data streams
            • A data stream created from a local file upload can't be automatically refreshed. You can upload a new file to refresh your data.
            		 Yes  
            Maximum number of data streams created from a local file upload
            • 100
            		 Yes  
            Unsupported file or column names
            • File or column names with multi-byte characters aren't supported.
            		 Yes  
            Primary key uniqueness restrictions
            • Primary key uniqueness isn’t enforced for local CSV file uploads.
            		 We recommend that your primary key column is unique before uploading a local file.

            SFTP Ingestion

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            Maximum size per file
            • 2 GB
            		 Yes The SFTP Connector can load up to 1,000 files that are less than 2 GB but total no more than 30 GB per data stream run. If you have more data, you can use upsert mode for consecutive runs.
            Maximum number of files included in a scheduled run
            • 1,000
            		 Yes
            Maximum data per data stream run
            • 30 GB
            		 Yes
            Maximum size limit for file allowed during field analysis
            • 4.1 MB
            		 Yes  
            Maximum file for PGP-encrypted schema file analysis
            • 25 MB
            		 Yes  
            Maximum object size for fields
            • 25 MB
            		 Yes  
            Maximum total size allowed for manual extract
            • 4.5 GB
            		    
            Maximum total number of files allowed for manual extract
            • 1,000
            		 Yes  

            Data Model Object Guidelines and Limits

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            Fields per data model object
            • 800 per field type
            • 1,050 total fields
            		 Yes Both custom and system fields are counted toward this limit.
            Maximum number of data model objects
            • 7,500
            • When using Data 360 in a Developer org: 300
            		 Yes  
            Maximum number of custom relationships per data model object
            • 25
            		 Yes  

            Data Shares Guidelines and Limits

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            Total number of data shares per org
            • 10
            		 Yes  
            Total number of data share targets per org
            • 20
            		 Yes This limit isn’t applicable for Salesforce Data 360 data share targets.
            Maximum number of objects included in a data share
            • 20
            		 Yes  

            Data Transforms Guidelines and Limits

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            Total number of streaming data transforms per org
            • 25
            		 Yes  
            Maximum number of characters in a streaming data transform SQL statement
            • 20,000
            		 Yes  
            Maximum number of batch transforms that can be created in an org
            • 1,000
            		 Yes Contact your Account Executive to adjust this limit.
            Maximum runtime (duration) of a batch transform job before it’s canceled by the system
            • 24 hours
            		 Yes  
            Maximum concurrent batch transforms that can be run at a time
            • 5
            		 Yes  

            AI Model Builder Guidelines and Limits

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            Connected Models (Bring Your Own Model (BYOM)): Maximum number of input variables for an AI model per tenant
            • 100
            		 Yes  
            Einstein-Created Models (Created From Scratch): Maximum number of input variables for an AI model per tenant
            • 50
            		 Yes  
            Einstein-Created Models (Created From Scratch): Numeric variables automatically categorized into groups
            • 10 (numeric fields are grouped into these ranges based on their distribution)
            		 Yes If there are less than 10 unique numeric fields, the numbers won't be grouped.
            Einstein-Created Models (Created From Scratch): Number of unique categories per model variable
            • 100
            		 Yes

            If a text variable has more than 100 unique categories, the top 100 (ranked by occurrence) will remain unique. Any remaining categories are grouped into an "Other" category. Null values in Date, Date Time, and Text variables are grouped into a separate "Null" category.

            Missing numeric values are placed in an "Unspecified" category.

            While the category limit is 100 based on data provided, it's possible to have 102 categories.
            Einstein-Created Models (Created From Scratch): Minimum number of occurrences required to consider a single text value as unique for model training
            • 25
            		 Yes  
            Einstein-Created Models (Created From Scratch): Minimum number of rows
            • 400
            		 Yes  
            Einstein-Created Models (Created From Scratch): Maximum number of rows
            • 20 million
            		 Yes For the XGBoost algorithm, we automatically use a representative sample of only 5 million rows from the 20 million provided.
            Einstein-Created Models (Created From Scratch): Maximum number of concurrent model training jobs per org
            • 2
            		   Any additional model training jobs remain queued until a job opens up.
            All Predictive Models: Maximum number of active prediction jobs per org
            • 100
            		    
            All Predictive Models: Maximum number of inference calls per month
            • 500 million
            		    
            All Predictive Models: Maximum number of active models
            • 5
            		    
            All Predictive Models: Maximum number of rows for each inference request to an external provider
            • 200
            		    

            Identity Resolution Guidelines and Limits

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            Total number of identity resolution rulesets
            • 2 per primary data model object per data space
            		 Yes

            This limit is enforced in the UI.

            Consolidating profiles using identity resolution reduces the number of Individual profiles counted against contracted entitlements. All unified profiles created from all rulesets are summed, so deleting unneeded rulesets reduces total unified profiles.

            Consolidating profiles using identity resolution impacts the number of known and anonymous profiles in your org. To understand how many anonymous profiles are allowed, review the Number of Anonymous Profiles limit.
            Maximum scheduled job frequency per ruleset
            • Once per day
            • When using Data 360 in a Developer org: 0
            		 Yes

            Because the time of day can vary, ruleset jobs can occasionally run more than once in a 24-hour period. Scheduled ruleset jobs are skipped if there are no changes to source data, object mappings, or ruleset configurations. This limit is enforced automatically and isn't configurable.

            When using Data 360 in a Developer org, ruleset runs aren’t scheduled. Use Run Ruleset to kick off a ruleset job manually.

            Maximum number of ruleset jobs in any 24-hour period
            • 4 per ruleset per data space
            		 Yes Ruleset jobs don’t run if there are no changes to source data, object mappings, or ruleset configurations, or if you’ve already run 4 ruleset jobs in the last 24 hours.
            Maximum number of match rules per ruleset
            • 10
            		 Yes  
            Maximum number of match criteria per match rule
            • 10
            		 Yes  
            Maximum number of source profiles that can be unified into a single unified profile
            • 50,000
            		 Yes Large profiles usually result from poor data quality or match rules that are too broad. To reduce the number of matching records, clean your data or refine your match rules.
            Maximum size of source records processed in a ruleset
            • 15 KB
            		 Yes Records that are larger than 15 KB are skipped when rulesets run.
            Maximum number of match rules per ruleset
            • 10
            		 Yes  
            Maximum number of match criteria per match rule
            • 10
            		 Yes  
            Maximum combined character count of values reviewed by a match rule
            • 500 characters
            		 Yes If the sum of characters of values being matched by a single match rule is greater than 500, some values will be truncated during matching. This can result in incorrect matches.
            Match methods used during real-time matching
            • Real-time identity resolution isn't available
            		 Yes  

            Segmentation Guidelines and Limits

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            Total number of active segments
            • 9,950
            • When using Data 360 in a Developer org: 25
            		 Yes An active segment is a created segment with all functionality available. These limits are enforced by the UI. Learn more in Create a Segment in Data 360 and Segment Types and Statuses.
            Total number of attributes with value suggestions
            • 500
            		 Yes Value suggestions can be enabled for up to 500 attributes for your entire org. Learn more in Use Value Suggestions in Segmentation.
            Total number of standard concurrent publishes
            • 50
            		 Yes  
            Total number of filters
            • 100
            		 Yes You can have up to 50 filters each in the Include and Exclude tabs. Learn more in Segment Your Data in Attributes.
            Total number of months that events are queried for standard segments
            • 24
            		 Yes The segmentation event limit is 24 months or less. This limit is enforced by the UI. For scheduled and manual publishes, segments error if a date range goes past the 24-month range. To run the segment again, adjust your event dates to fit that range.
            Maximum number of scheduled publishes per standard segment per day
            • 2
            • When using Data 360 in a Developer org: 0
            		 Yes

            This limit is enforced by the UI. Learn more in

            Publish a Segment in Data 360.

            When using Data 360 in a Developer org, segment publishes aren’t scheduled. Use Publish Now to kick off publication manually.
            Maximum number of rapid segments
            • 20
            		    
            Total number of rapid concurrent publishes
            • 20
            		    
            Total number of days that events are queried for rapid segments
            • 7
            		   The segmentation event limit is 7 days. This limit is enforced by the UI. For scheduled and manual publishes, segments error if a date range goes past the 7-day range. To run the segment again, adjust your event dates to fit that range.
            Maximum number of scheduled publishes per rapid segment per day
            • 24
            		    
            Real-time segments
            • Not available
            		 Yes  
            Total number of processed records for rapid segments
            • 600 million
            		 Yes This limit is for records processed for a rapid segment, which is the sum of records present in the data streams associated with the segment. This limit isn’t associated with the segment population.
            Maximum number of Einstein segments created per month
            • 160
            		 Yes  

            Streaming Insights Guidelines and Limits

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            Aggregation time window
            • Minimum: 1 minute
            • Maximum: 24 hours
            		 Yes Learn more about the differences between Streaming Insights and Calculated Insights.
            Total number of dimensions
            • 10
            		 Yes  
            Total number of measures
            • 5
            		 Yes  
            Total number of Streaming Insights
            • 5
            		 Yes You can create more than 5 streaming insights. However, only the first 5 streaming insights run.
            Unsupported features
            • Streaming insights aren’t available in segmentation or activation
            		 Yes  
            Supported primary DMO objects
            • DMOs with specific streaming sources
            		  

            Primary objects are filtered DMOs created from any of these streaming connectors:

            • Web SDK
            • Mobile SDK
            • Marketing Cloud Personalization
            • Ingestion API (Streaming)
            • Ingestion API (Batch)
            Allowed joins
            • Individual DMO and unified objects
            		  

            During the identity resolution process, unified objects from the Individual DMO include these objects:

            • Unified Individual
            • Individual Link
            • Contact Point

            Joins on objects unrelated to Individual or objects related to Account aren't allowed.

            Unstructured Data and Search Index Guidelines and Limits

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            Unstructured data and search indexes
            • Not available
            		 Yes  

            API Guidelines and Limits

            The concurrent request and query limits of Data 360 APIs are independent from the concurrency limits governing the Salesforce platform APIs.

            APIs: Query, Insights, and Profile

            There are three distinct classes of APIs used to extract data: Profile, Query, and Calculated Insights. These API limits for CDP orgs are based on the org’s Data 360 edition, not its Salesforce edition, except when running Data 360 on a Developer org.

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            Maximum timeout
            • 5 minutes
            		 Yes The HTTP 408 status code indicates a timeout.
            Total number of queries per day for Query API and Insights API
            • Corporate Edition: 1,000
            • Enterprise and Enterprise Editions Plus: 1,000
            		   Applies to Query API v1 and v2.
            Total number of concurrent inbound requests for Insights API
            • Corporate Edition: 15
            • Enterprise and Enterprise Editions Plus: 15
            		    
            Total number of queries per day for Profile API
            • Corporate Edition: 10,000
            • Enterprise Edition: 50,000
            		    
            Total number of queries per month for Profile API
            • Corporate Edition: 100,000
            • Enterprise Edition: 500,000
            		    

            Ingest Bulk API

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            API usage limits
            • Varies
            		 Yes After each request, your app must check the response code. The HTTP 429 Too Many Requests status code indicates that the app must reduce its request frequency.
            Bulk job retention time
            • 7 days
            		 Yes Open bulk jobs with the status of Open or Upload Complete that are older than 7 days are deleted from the ingestion queue.
            Maximum number of files per job
            • 100 files
            		    
            Maximum payload size
            • 150 MB
            		 Yes Referring to the CSV files uploaded via Bulk API.
            Number of requests or jobs allowed per hour
            • 20
            		    
            Number of concurrent requests or jobs Allowed at One Time
            • 5
            		    

            Ingest Streaming API

            These limits and guidelines are for streaming ingestion across Mobile and Web SDK and Ingestion API.

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            API usage limits
            • Varies
            		 Yes After each request, your app must check the response code. The HTTP 429 Too Many Requests status code indicates that the app must reduce its request frequency.
            Expected latency for Ingestion API
            • Real-time: 300 milliseconds
            • Non-real time: 3 minutes

            Data is processed asynchronously approximately every 3 minutes.

            The latency means that received records are processed every two minutes, but depending on the data volume and how busy the scheduler is, it can take additional time until the data is committed to storage and available for consumption.
            Expected latency for Mobile and Web SDK and server-to-server applications
            • Profile data: up to 3 minutes
            • Engagement data: up to 2 minutes
            • Real-time for Profile and Engagement data: 300 milliseconds
            		 To reduce processing power and bandwidth requirements, mobile events are queued on the device.
            Maximum number of records that can be deleted via Ingestion API deletion
            • 200 records
            		 Yes  
            Maximum payload size per request
            • 200 KB
            		 Yes JSON data uploaded via Streaming API have a maximum body size of 200 KB per request. The HTTP 403 Forbidden status code indicates that the API request exceeded the 200 KB limit.
            Total number of requests per second across all Ingestion API object endpoints
            • 250 requests
            		 Yes The HTTP 429 Too Many Requests status code indicates that the app has exceeded the 250-request limit and must reduce its request frequency.

            Profile API

            Feature or Function Guidelines and Limits Hard Limit? Additional Information and Resources
            Maximum number of records returned per call
            • 49,999
            		 Yes  
            Total number of fields per record
            • 50
            		 Yes  
            Note
            Note Unified Profile objects that are filtered on the index column have accelerated response rates. Other Profile category objects can’t guarantee a faster response.
            .
             
            Loading
            Salesforce Help | Article