Awss3 Receiver
contrib
Maintainers: @atoulme, @adcharre
Source: opentelemetry-collector-contrib
Supported Telemetry
Overview
Overview
Receiver for retrieving telemetry data (traces, metrics, and logs) previously stored in S3. This receiver is commonly used with the AWS S3 Exporter but can process any S3 objects containing telemetry data in supported formats. The receiver supports:- Traces, Metrics, and Logs - All three OpenTelemetry signal types
- Multiple formats - OTLP (JSON and Protocol Buffers), plus custom formats via encoding extensions
- Two retrieval modes:
- Time-based: Fetch data from a specific time range using S3 key partitioning
- Event-driven: Process new objects as they arrive via SQS notifications
Configuration
The following exporter configuration parameters are supported.| Name | Description | Default | Required |
|---|---|---|---|
starttime | The time at which to start retrieving data. | Required if fetching by time | |
endtime | The time at which to stop retrieving data. | Required if fetching by time | |
s3downloader: | |||
region | AWS region. | ”us-east-1” | Optional |
s3_bucket | S3 bucket | Required | |
s3_prefix | prefix for the S3 key (root directory inside bucket). | Required | |
s3_partition_format | Format for the partition key, See strftime for format specification. | "year=%Y/month=%m/day=%d/hour=%H/minute=%M" | Optional |
s3_partition_timezone | IANA timezone name applied when formatting the partition key. | Local time | Optional |
file_prefix | file prefix defined by user | Optional | |
file_prefix_include_telemetry_type | whether to append <telemetry_type>_ to the file prefix when building S3 keys | true | Optional |
endpoint | overrides the endpoint used by the exporter instead of constructing it from region and s3_bucket | Optional | |
endpoint_partition_id | partition id to use if endpoint is specified. | ”aws” | Optional |
s3_force_path_style | set this to true to force the request to use path-style addressing | false | Optional |
sqs: | |||
queue_url | The URL of the SQS queue that receives S3 bucket notifications | Required if fetching by SQS notification | |
region | AWS region of the SQS queue | Required if fetching by SQS notification | |
endpoint | Custom endpoint for the SQS service | Optional | |
max_number_of_messages | Maximum number of messages to retrieve in a single SQS request | 10 | Optional |
wait_time_seconds | Wait time in seconds for long polling SQS requests | 20 | Optional |
encodings: | An array of entries with the following properties: | Optional | |
extension | Extension to use for decoding a key with a matching suffix. | Required | |
suffix | Key suffix to match against. | Required | |
notifications: | |||
opampextension | Name of the OpAMP Extension to use to send ingest progress notifications. | ||
tag_object_after_ingestion | If enabled the receiver will attempt to tag the object after successfully ingesting it. | false | Optional |
- Time Range Mode - Specify
starttimeandendtimeto fetch data from a specific time range. - SQS Message Mode - Subscribe to SQS messages to process new objects as they arrive.
SQS Message Configuration
The receiver can subscribe to an SQS queue that receives S3 event notifications:starttime/endtime) and SQS configuration cannot be used together.
Time format for starttime and endtime
The starttime and endtime fields are used to specify the time range for which to retrieve data.
The time format is either RFC3339,YYYY-MM-DD HH:MM or simply YYYY-MM-DD, in which case the time is assumed to be 00:00.
Encodings
By default, the receiver understands the following encodings:- otlp_json (OpenTelemetry Protocol format represented as json) with a suffix of
.json - otlp_proto (OpenTelemetry Protocol format represented as Protocol Buffers) with a suffix of
.binpb
encodings options allows you to specify Encoding Extensions to use to decode keys with matching suffixes.
Example Configuration
Notifications
The receiver can send notifications of ingest progress to an OpAmp server using the custom message capability of “org.opentelemetry.collector.receiver.awss3” and message type “TimeBasedIngestStatus”. The format of the notifications is a ProtoBuf formatted OLTP logs message with a single Log Record. Thebody of the
record is set to status and the timestamp of the record is used to hold the ingest time. The record also has the
following attributes:
| Attribute | Description |
|---|---|
telemetry_type | The type of telemetry being ingested. One of “traces”, “metrics”, or “logs”. |
ingest_status | The status of the data ingestion. One of “ingesting”, “failed”, or “completed”. |
start_time | The time to start retrieving data as an Int64, nanoseconds since Unix epoch. |
end_time | The time to stop retrieving data as an Int64, nanoseconds since Unix epoch. |
failure_message | Error message if ingest_status is “failed”. |
ingest_status set to “failed” status with
the time of the data being ingested when the failure occurred.
If the ingest process completes successfully a status message with ingest_status set to “completed” is sent.
Object Lifecycle Management
If thetag_object_after_ingestion is enabled the receiver will make a best-effort attempt to tag objects with otel-collector:status = ingested after they are processed by the pipeline. This requires an additional s3:PutObjectTagging permission.
This tag can then be used with a lifecycle policy to expire ingested objects or transition them to cheaper storage classes.
Configuration
Example Configuration
Last generated: 2026-04-13