Bulk API Limits | Bulk API Developer Guide

Bulk API use is subject to the standard API usage limits. Each HTTP request counts as one call for the purposes of calculating usage limits.

Each batch must contain exactly one CSV, XML, or JSON file containing records for a single object, or the batch is not processed and stateMessage is updated. Use the Enterprise WSDL for the correct format for object records.

You can submit up to 15,000 batches per rolling 24-hour period. You can’t create batches associated with a job that is more than 24 hours old. If a batch is submitted to a closed job, the batch will not be created, however it will still count against the batch allocation as a submitted batch.

Batches and jobs that are older than 7 days are removed from the queue if they are in a terminal state (completed or failed). The seven days are measured from the youngest batch associated with a job, or the age of the job if there are no batches. You can’t create batches associated with a job that is more than 24 hours old.

Batches for data loads can consist of a single CSV, XML, or JSON file that is no larger than 10 MB.
A batch must contain some content or an error occurs.

For binary content limits, see General Limits.

Batches are processed in chunks. The chunk size depends on the API version. In API version 20.0 and earlier, the chunk size is 100 records. In API version 21.0 and later, the chunk size is 200 records. There’s a 5-minute limit for processing each chunk. Also, if it takes longer than 10 minutes to process a whole batch, the Bulk API places the remainder of the batch back in the queue for later processing. If the Bulk API continues to exceed the 10-minute limit on subsequent attempts, the batch is placed back in the queue and reprocessed up to 10 times before the batch is permanently marked as failed.

Even if the batch failed, some records could have completed successfully. If you get a timeout error when processing a batch, split your batch into smaller batches, and try again.

Bulk queries have an extra 2-minute limit for processing the query, which is separate from the batch processing limit.

Note

The only valid compression value is gzip. Compression is optional, but recommended. Compression doesn’t affect the character limits defined in Batch size.

Any user with correct permission can abort a job. Only the user who created a job can close it.

Only the user who created a job can close it. Any user with correct permission can abort a job.

Each job can specify one operation and one object. Batches associated with this job contain records of one object. Optionally, you can specify serial processing mode, which is used only when previously submitted asynchronous jobs have accidentally produced contention because of locks. Use only when advised by Salesforce.

You can't edit the value of an external ID field in JobInfo. When specifying an external ID, the operation must be upsert. If you try to use it with create or update, an error is generated.

Batches and jobs that are older than 7 days are removed from the queue if they are in a terminal state (completed or failed). The seven days are measured from the youngest batch associated with a job, or the age of the job if there are no batches. You can’t create batches associated with a job that is more than 24 hours old.

The maximum time that a job can remain open is 24 hours. The Bulk API doesn't support clients that, for example, post one batch every hour for many hours.

After a job has completed, the job status and batch result sets are available for seven days after which this data is deleted permanently.

When you submit a POST body with a change in job status, you can only specify the status field value. If operation or entity field values are specified, an error occurs.

Regardless of whether the “API Enabled” profile permission is granted, portal users (Customer Portal, Self-Service portal, and Partner Portal) can't access Bulk API.

Bulk API does not support queries with any of the following:

GROUP BY, OFFSET, or TYPEOF clauses
Aggregate functions such as COUNT()
Date functions in GROUP BY clauses (date functions in WHERE clauses are supported)
Compound address fields or compound geolocations fields

Item	Bulk API Limit
Maximum number of records uploaded per rolling 24-hour period	150,000,000 (15,000 batches x 10,000 records per batch maximum)
Maximum time before a batch is retried	10 minutes
Maximum file size	10 MB
Maximum number of characters in a field	32,000
Maximum number of fields in a record	5,000
Maximum number of characters in a record	400,000
Maximum number of records in a batch	10,000
Maximum number of characters for all the data in a batch	10,000,000

Item	Bulk API Limit
Number of attempts to query	15 attempts at 10 minutes each to process the batch. There is also a 2-minute limit on the time to process the query. If more than 15 attempts are made for the query, an error message of “Tried more than fifteen times” is returned. If the query takes more than 2 minutes to process, a QUERY_TIMEOUT error is returned.
Number of retrieved files	15 files. If the query returns more than 15 files, add filters to the query to return less data. Bulk batch sizes are not used for bulk queries.
Results lifespan	7 days
Maximum retrieved file size	1 GB
Number of batches that can be submitted per 24-hour rolling window	You can submit up to 15,000 batches per rolling 24-hour period. This allocation is shared between Bulk API and Bulk API 2.0, so every batch that is processed in Bulk API or Bulk API 2.0 counts towards this allocation.
Total query results that can be generated per 24 hour rolling window	N/A

Item	Bulk API Limit
Batch and job lifespan	Batches and jobs that are older than seven days are removed from the queue if they are in a terminal state (completed or failed) regardless of job status. The seven days are measured from the youngest batch associated with a job, or the age of the job if there are no batches. You can’t create batches associated with a job that is more than 24 hours old.
Batch processing time	Batches are processed in chunks. The chunk size depends on the API version. In API version 20.0 and earlier, the chunk size is 100 records. In API version 21.0 and later, the chunk size is 200 records. There’s a 5-minute limit for processing each chunk. Also, if it takes longer than 10 minutes to process a whole batch, the Bulk API places the remainder of the batch back in the queue for later processing. If the Bulk API continues to exceed the 10-minute limit on subsequent attempts, the batch is placed back in the queue and reprocessed up to 10 times before the batch is permanently marked as failed.
Binary content	The length of any file name can’t exceed 512 bytes. A zip file can’t exceed 10 MB. The total size of the unzipped content can’t exceed 20 MB. A maximum of 1,000 files can be contained in a zip file. Directories don’t count toward this total.
Maximum time that a job can remain open	24 hours.

For a comparison with Bulk API 2.0 limits, see the Salesforce Developer Limits and Allocations Quick Reference.