# OTLP Endpoints
OpenTelemetry Protocol (OTLP) is a general-purpose telemetry data delivery protocol designed for the OpenTelemetry. CloudWatch OpenTelemetry endpoints are HTTP 1.1 endpoints. You need to configure your OpenTelemetry collector to start sending open telemetry data to CloudWatch. For more information, see [Getting started](CloudWatch-OTLPGettingStarted.md).
## Logs endpoint
The logs endpoint follows the pattern `https://logs.AWS Region.amazonaws.com/v1/logs`. For example, for the `US West (Oregon) (us-west-2) Region`, the endpoint will be `https://logs.us-west-2.amazonaws.com/v1/logs`. You can use the above endpoint to forward logs to an existing `LogGroup` and `LogStream`. For more information on setting up `LogGroup` to ingest log data, see [Amazon CloudWatch Logs concepts](https://docs.aws.amazon.com/AmazonCloudWatch/latest/logs/CloudWatchLogsConcepts.html).
You must configure `LogGroup` and `LogStream` when you invoke CloudWatch Logs OpenTelemetry endpoint by setting `x-aws-log-group` and `x-aws-log-stream` HTTP headers to `LogGroup` and `LogStream` name respectively. For more information, see [Getting started](CloudWatch-OTLPGettingStarted.md). The endpoint authenticates callers using Signature 4 authentication. For more information, see [AWS Signature Version 4 for API requests](https://docs.aws.amazon.com/IAM/latest/UserGuide/reference_sigv.html).
When log event size exceed 1MB, CloudWatch Logs automatically truncates up to 10 fields, starting with the largest fields. Each field is truncated as needed to keep the total event size as close to 1MB as possible. The excess portions are stored as Large Log Objects (LLOs) and LLO reference system fields are added. Optionally, you can specify the field paths that need to be truncated by setting the `x-aws-truncatable-fields `HTTP header. The LLOs can be retrieved and streamed back using the `GetLogObject` API. For more information, see [GetLogObject](https://docs.aws.amazon.com/AmazonCloudWatchLogs/latest/APIReference/API_GetLogObject.html). Support of log events larger than 1MB and LLO experience is available in US East (N. Virginia) US West (Oregon), Europe (Frankfurt), Asia Pacific (Sydney), Asia Pacific (Mumbai), US East (Ohio), Europe (Ireland), Asia Pacific (Tokyo), and Asia Pacific (Singapore).
## Metrics endpoint
The metrics endpoint follows the pattern `https://monitoring.AWS Region.amazonaws.com/v1/metrics`. For example, for the US West (Oregon) (us-west-2) Region, the endpoint will be `https://monitoring.us-west-2.amazonaws.com/v1/metrics`.
You need to configure your OpenTelemetry collector to start sending metrics to CloudWatch. The endpoint authenticates callers using Signature 4 authentication. For more information, see [AWS Signature Version 4 for API requests](https://docs.aws.amazon.com/IAM/latest/UserGuide/reference_sigv.html).
## Traces endpoint
The traces endpoint follows the pattern `https://xray.AWS Region.amazonaws.com/v1/traces`. For example, for the US West (Oregon) (us-west-2) Region, the endpoint will be `https://xray.us-west-2.amazonaws.com/v1/traces`.
You need to configure your OpenTelemetry collector to start sending traces to CloudWatch. The endpoint authenticates callers using Signature 4 authentication. For more information, see [AWS Signature Version 4 for API requests](https://docs.aws.amazon.com/IAM/latest/UserGuide/reference_sigv.html).
## RUM endpoint
The RUM endpoint follows the pattern `https://dataplane.rum.{AWS Region}.amazonaws.com/v1/rum`. For example, for the US West (Oregon) Region, the endpoint is `https://dataplane.rum.us-west-2.amazonaws.com/v1/rum`. This endpoint handles *client-side telemetry data* (only traces and log records with `eventName`) for CloudWatch RUM applications.
To use this endpoint, you must create a [RUM app monitor](CloudWatch-RUM-get-started-create-app-monitor.md) with Mobile platform (Android/ iOS) and use the code snippet generated to instrument your applications. The snippet will pull the RUM Mobile SDKs which are configured with this endpoint. You can configure the SDKs further for RUM to collect telemetry accordingly.
The endpoint supports both authenticated and unauthenticated requests. You can use AWS Signature Version 4 (SigV4) for authenticated requests, or resource-based policies to allow unauthenticated access from mobile applications.
To learn more about the authentication models as defined in their SDKs, see the following:
+ iOS applications - [AWS Distro for OpenTelemetry (ADOT) iOS SDK](https://github.com/aws-observability/aws-otel-swift).
+ Android applications - [AWS Distro for OpenTelemetry (ADOT) Android SDK](https://github.com/aws-observability/aws-otel-android).
## Endpoint limits and restrictions
The table lists the common endpoint limits and restrictions for metrics, traces, and logs.
| Limit | Endpoint | Additional information |
| --- | --- | --- |
| Required collector extension | [sigv4authextension](https://github.com/open-telemetry/opentelemetry-collector-contrib/tree/main/extension/sigv4authextension) | To send metrics or traces to OTLP endpoint you must use sigv4authextension |
| Supported protocol | HTTP | The endpoint supports only HTTP and doesn't support gRPC |
| Supported OTLP versions | OTLP 1.x | |
| Payload format | binary, json | The endpoint accepts requests using binary and json formats |
| Compression methods | gzip, none | The endpoint only supports gzip and none compression methods |
The table lists the endpoint limits and restrictions for traces.
| Limit | Traces endpoint | Additional information |
| --- | --- | --- |
| Maximum uncompressed bytes / request | 5 MB | The OTLP endpoint will reject requests larger than 5MB when payload is uncompressed. |
| Maximum events / request | 10,000 spans | The maximum number of spans in a batch is 10,000. Exceeding this limit will cause rejection of the API call. |
| Single resource and scope size | 16 KB | Each unique resource and corresponding scope should not exceed 16 KB of size. Exceeding this limit for any resource will cause rejection of the entire API call. |
| Single span maximum size | 200 KB | Spans more than 200KB will be rejected by the endpoint. |
| Span created timestamps | 2 hours in the future and 14 days in the past | None of the spans in the batch can be more than two hours in the future or more than 14 days in the past. |
| Maximum time gap in events / request | 24 hours | |
The table lists the endpoint limits and restrictions for metrics.
| Limit | Metrics endpoint | Additional information | Error code |
| --- | --- | --- | --- |
| Maximum TPS | 500 | The maximum number of requests per second allowed per account. | 429 |
| Maximum new series creation rate | 1,000,000 | Maximum number of new series that can be created in a 10-minute window. This limit applies only to creating new metric names, not to ingesting data points for existing metrics. | 429 - all metrics in the requests are throttled; 200 - metrics in the requests are partially throttled |
| Maximum request size | 1 MB | The maximum uncompressed size of the request. | 400 |
| Maximum datapoint count | 1,000 | The maximum number of datapoints sent in a single request. This number is a sum across ResourceMetrics, ScopeMetrics and Metrics. | 400 |
| Maximum metadata size | 40 KB | The maximum combined size of all labels and label values accepted for a series per datapoint. | 400 - all metrics in the requests are invalid; 200 - metrics in the requests are partially invalid |
| Maximum label count | 150 | Maximum number of labels across Resource/Scope/Datapoint attributes per datapoint. | 400 - all metrics in the requests are invalid; 200 - metrics in the requests are partially invalid |
The table lists the endpoint limits and restrictions for logs.
[\[See the AWS documentation website for more details\]](http://docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/CloudWatch-OTLPEndpoint.html)
**Note**
The account limits for Logs are shared across the SDK and the new Logs endpoint.