mParticle imposes certain limits on incoming data in order to protect the performance of both the mParticle dashboard and your apps. This includes limits around the length of individual data points, such as event names, how fast mParticle can receive data, and how many unique data points a workspace or account can have.
The tables below list mParticle’s Default Limits. Enterprise customers should contact their Customer Success Manager for custom limits.
mParticle can receive data across many channels, and limits are not always enforced in the same way for each channel. Where appropriate, the details section of each table describes how limits affect SDK data - received from mParticle’s native SDKs - and S2S or ‘server-to-server’ data. S2S data includes data received via the Events API, and from partner feeds.
The following limits apply to event batches being ingested.
Resource |
Limits | Details |
|---|---|---|
| Event name length | 256 characters | SDK logs an error and the event is not uploaded. For S2S data, names exceeding the limit is truncated. |
| Event attribute name length | 256 characters | SDK logs an error and no attributes are set for the event. S2S truncates the attribute name. |
| Event attribute value length | 4096 characters | SDK logs an error and no attributes are set for the event. |
| User attributes per event batch | 100 | SDK allows only 100 attributes per user and logs an error if you attempt to create additional attributes. |
| User attribute name length | 256 characters | SDK logs an error and no attributes are set. |
| User attribute value length | 4096 characters | SDK logs an error and no attributes are set. |
| User attribute list | 1000 Entries | SDK logs an error and no attributes are set. |
| User attribute list entry length | 512 characters | SDK logs an error and no attributes are set. |
The following limit applies only to data being ingested with an SDK.
Resource |
Limits | Details |
|---|---|---|
| Total batch size | 128KB | If this limit is exceeded, the SDK automatically attempts to break up the batch into multiple smaller batches. |
The following limits apply to events per workspace or user.
Resource |
Limits | Details |
|---|---|---|
| Unique event names and Screen Names per workspace | 1000 | New unique event names over the limit are dropped from incoming data. You can request this limit be raised, but proliferating unique event names usually indicate problems with your data and can impact performance of both the mParticle dashboard and your apps. Therefore, the limit won’t be raised except where absolutely necessary. |
| Average events per user within 24 hours | 150 | This limit can be raised by request. |
| Average events per user within 30 days | 175 | This limit can be raised by request. |
mParticle reserves the right to restrict average events per user to ensure platform quality of service.
The following limits apply to resources in the Events API.
Resource |
Limits | Details |
|---|---|---|
| Batches per second | Variable but defaults to 270 batches per second | If exceeded, the Events API returns a 429 HTTP response code. Learn more about how the Events API is throttled in API Throttling. |
| Total Request Size | 256kb | Whether using the /events or /bulkevents endpoint, the total request size must be under 256kb regardless of how many batches are included in the request. |
The following limits apply to resources in the Platform API.
| Resource | Limits | Details |
|---|---|---|
| Data points per filter update | 500 | Actual limits scale up and down with demand. If you exceed this limit, the mParticle API could take more than 30s to respond, or time out. Set up your implementation to respect 504 responses and retry your request with smaller batches. Alternatively, you can use the UI as outlined in this guide. |
The following limit applies to resources in the Profile API.
Resource |
Limits | Details |
|---|---|---|
| Requests per second | Variable but starting at 15 requests per second | Actual limits scale up and down with demand. If you exceed the limit, the mParticle API returns an HTTP 429 response code. Set up your implementation to respect 429 responses and retry the request in an exponential backoff pattern. |
The following limits are explained in detail in Data Retention.
Resource |
Limits | Details |
|---|---|---|
| Event batch long-term archival storage | 24 months | Contact your mParticle Customer Service representative if you need longer or shorter archival storage. |
| Profile storage | 30 days | User Profiles are deleted after 30 days of inactivity. |
| Real-time audience storage | 30MB | Maximum size of a single users data in real-time audience storage. Typical users are ~200kb. |
The following limits apply to resources in the Data Subject Request API.
Resource |
Limits | Details |
|---|---|---|
| Requests per minute | 1000 requests per minute | This limit applies across three API actions (GET, POST, and DELETE) and is enforced per Workspace |
| Total new requests | 80,000 requests per day | This limit applies to the POST API action and is enforced per Workspace. It is split into sliding intervals of 8 hours, meaning that a maximum of 26,666 requests are allowed in any 8 hour window. |
The following limits apply to resources used in data planning.
The following limits apply to resources in the Data Planning API.
Resource |
Limits | Details |
|---|---|---|
| Requests per minute per account | 3000 requests per minute | This limit applies to all GET, POST, and PATCH API actions. |
| Requests per minute per organization | 6000 requests per minute | This limit applies to all GET, POST, and PATCH API actions. |
The following limits apply to resources in the Warehouse Sync API.
Limit |
Value | Notes |
|---|---|---|
| Max number of Active Pipelines | 5 | |
| Catchup limit for new hourly pipeline | 7 days | |
| Catchup limit for new daily pipeline | 6 months | |
| Catchup limit for new weekly pipeline | 3 years | |
| Catchup limit for new monthly pipeline | 5 years | |
| Record count limit per hourly interval | 1 million | |
| Record count limit per daily interval | 24 million | |
| Record count limit per weekly interval | 40 million | |
| Record count limit per monthly interval | 40 million | |
| Record count limit per once request | 40 million | |
| Record count limit per on-demand request | 24 million | This request must be used with an on-demand trigger. |
The following limits apply to workspaces, users, and the name length limits for audiences and tags.
Resource |
Limit | Details |
|---|---|---|
| Workspaces | 50 | Users are prevented from creating additional workspaces. This limit can be raised by arrangement. |
| Users | 200 | Admins are prevented from creating additional users. This limit can be raised by arrangement. |
| Audience name Length | 100 | Audience name and external name fields are limited to 100 characters. |
| Tag name length | 18 | Tag names are limited to 18 characters. |
mParticle APIs have two types of rate limits in place to protect mParticle’s servers from high demand:
You can configure your mParticle integration to respond programmatically to prevent data loss according to the 429 response header you receive.
Throttled resources include a percentage representation of your current total usage in 2xx response headers. These percentage-used headers allow you to modify your request speed or acceleration to prevent exceeding a rate limit.
An API request that exceeds the rate (speed) limit of a resource receives a 429 response and a header with the format:
X-mp-rate-limit-exceeded: "LIMIT"
where "LIMIT" is the ID of the source of the limit as described in the table below:
| Value | Description |
|---|---|
"org" |
A limit applied to an entire organization. |
"account" |
A limit applied to an account ID within a parent organization. |
"app" |
A limit applied to an app ID within a parent organization. |
"feed" |
A limit applied to a feed ID within a parent organization. |
"workspace" |
A limit applied to a workspace ID within a parent organization. |
"system" |
A limit applied at an mParticle system level across multiple organizations. |
"acceleration" |
An acceleration limit that can be applied to any of the speed limits listed above. |
An API request that exceeds any applicable acceleration limit receives a 429 response and a header with the format:
X-mp-rate-limit-exceeded: “acceleration”
Rate limited endpoints return an X-mp-rate-limit-percentage-used header with 2xx responses including the percentage of the limit used. If a resource has multiple limits, the response header includes the greatest consumption percentage of all applicable speed and acceleration limits, omitting any user-based limits.
For example, if a request is not throttled and is subject to both a speed and acceleration limit, and the current consumption is 92% of the speed limit and 50% of the acceleration limit, then the header lists the consumption of the speed limit because it is higher than the current acceleration limit usage:
X-mp-rate-limit-percentage-used: 92
If you receive a 429 response for exceeding a speed limit, reduce the frequency of your requests. Use exponential back-off with jitter and respect the RetryAfter value which is a non-negative decimal integer indicating the number of seconds to delay your request.
If you receive a 429 response for exceeding an acceleration limit, you can still submit requests but you should slow the increase of your request speed. Use exponential back-off with jitter when determining the new frequency of your requests.
Was this page helpful?