feat(boto3): Add S3 service extension by pabloDeputter · Pull Request #7617 · getsentry/sentry-python

pabloDeputter · 2026-09-22T14:49:31Z

Description

add the following service-specific attributes:
- aws.s3.bucket required in almost all s3 operations except list-buckets.
- http.response.body.size which is the number of bytes in the payload.
add the following OTel attributes that are NOT yet in sentry-conventions:
- aws.s3.copy_source
- aws.s3.delete
- aws.s3.key
- aws.s3.part_number
- aws.s3.upload_id
- aws.s3.object_size (not in OTel) which is the total size of the s3 object itself.
  - extracted from ContentLength on GetObject and HeadObject responses; ObjectSize on GetObjectAttributes, and ContentLength on PutObject requests. For ranged GetObject or HeadObject calls, ContentLength is the used size.
more information where these parameters can be found/extracted from in https://opentelemetry.io/docs/specs/semconv/object-stores/s3/ & https://opentelemetry.io/docs/specs/semconv/registry/attributes/aws/#amazon-s3-attributes

Issues

Resolves #7576

github-actions · 2026-09-22T15:08:54Z

Codecov Results 📊

📊 Comparison with Base Branch

Metric	Change
Total Tests	📈 +621
Passed Tests	📈 +593
Failed Tests	📈 +1
Skipped Tests	📈 +27

➕ New Tests (1)

View new tests

test_cache_spans_item_size[True]
- File: tests.integrations.django.test_cache_module
- Status: ❌ Failing

❌ Failed Tests

`test_cache_spans_item_size[True]`

File: tests.integrations.django.test_cache_module
Suite: py3.12-django-v4.2.30
Error: tests/integrations/django/test_cache_module.py:763: in test_cache_spans_item_size assert not spans[0]["attributes"]["cache.hit"] E assert not True

Stack Trace

tests/integrations/django/test_cache_module.py:763: in test_cache_spans_item_size
    assert not spans[0]["attributes"]["cache.hit"]
E   assert not True

✅ Patch coverage is 83.12%. Project has 2569 uncovered lines.
✅ Project coverage is 90.35%. Comparing base (91c9bdb) to head (7c39a6c).

Files with missing lines (2)

File	Patch %	Lines
sentry_sdk/integrations/boto3/_services/s3.py	72.09%	⚠️ 12 Missing and 5 partials
sentry_sdk/integrations/boto3/_services/_attribute_extraction.py	93.75%	⚠️ 1 Missing and 1 partials

Coverage diff

@@            Coverage Diff             @@
##        master       #PR       +/-##
==========================================
+ Coverage    90.29%    90.35%    +0.06%
==========================================
  Files          195       204        +9
  Lines        26252     26625      +373
  Branches      9792      9882       +90
==========================================
+ Hits         23703     24056      +353
- Misses        2549      2569       +20
- Partials      1486      1493        +7

Generated by Codecov Action

sentry-warden · 2026-09-29T08:43:18Z

+        return {}
+
+    attributes = {}
+    for param, attribute, convert in specs:


PutObject object size is read from an absent response field

PutObject responses do not include Size, so this mapping leaves aws.s3.object_size unset. Extract ContentLength from the request params for PutObject, as the PR description specifies.

Evidence

_RESPONSE_OBJECT_SIZE_FIELDS maps PutObject to Size.

get_response_attributes() reads that field from the response and only sets aws.s3.object_size when it is an integer.

The S3 PutObject response has no Size; request extraction does not otherwise capture ContentLength.

Also found at 2 additional locations

sentry_sdk/integrations/boto3/_services/registry.py:15-15

sentry_sdk/integrations/boto3/_services/s3.py:28-31

_{Identified by Warden · code-review, find-bugs · F8T-EAX}

sentry-warden · 2026-09-29T08:43:18Z

+_SERVICE_EXTENSIONS: "Dict[str, Tuple[str, str]]" = {
+    "s3": ("sentry_sdk.integrations.boto3._services.s3", "_S3Extension"),
+}


S3 attribute extraction is not covered by tests

Please add tests that assert the new S3 request and response attributes, including the operation-specific object-size and response-body-size behavior. The existing S3 tests do not verify these fields, so regressions in the new extraction logic could go undetected.

Evidence

_S3Extension.get_request_attributes extracts S3 request fields and handles CompleteMultipartUpload object size.

_S3Extension.get_response_attributes has operation-specific logic for response body size and object size, including ranged HeadObject calls.

The boto3 tests do not assert SPANDATA.AWS_S3_* or HTTP_RESPONSE_BODY_SIZE; existing ApproxDict assertions allow additional attributes without checking them.

Also found at 1 additional location

sentry_sdk/integrations/boto3/_services/s3.py:64-109

_{Identified by Warden · code-review · VBM-G26}

sentry-warden · 2026-09-29T08:43:18Z

+    # the appropriate type.
+    # https://opentelemetry.io/docs/specs/semconv/object-stores/s3/
+    ("Bucket", SPANDATA.AWS_S3_BUCKET, _as_string),
+    ("CopySource", SPANDATA.AWS_S3_COPY_SOURCE, _as_string),


CopySource dict form silently drops aws.s3.copy_source

Boto3 often passes CopySource as a dict (Bucket/Key/VersionId); _as_string rejects that and omits aws.s3.copy_source—normalize dicts to the bucket/key string form.

Evidence

_REQUEST_ATTRIBUTES converts CopySource with _as_string.

_as_string only accepts non-empty str values and returns None for dicts.

Boto3 documents CopySource as str or dict; dict form is common for copy_object/copy_from, so the attribute is dropped on a frequent path.

Also found at 1 additional location

sentry_sdk/integrations/boto3/_services/registry.py:15-15

_{Identified by Warden · code-review, find-bugs · 8LX-QYJ}

sentry-warden · 2026-09-29T08:43:26Z

+
+        if operation_name in _RESPONSE_BODY_SIZE_OPERATIONS:
+            # `ContentLength` is the size of the HTTP body returned, which may be a range.
+            content_length = _as_integer(response.get("ContentLength"))
+            if content_length is not None and content_length >= 0:
+                attributes[SPANDATA.HTTP_RESPONSE_BODY_SIZE] = content_length
+
+        # these fields report the total S3 object size, not the HTTP body size.
+        object_size_field = _RESPONSE_OBJECT_SIZE_FIELDS.get(operation_name)
+        if object_size_field is not None:
+            object_size = _as_integer(response.get(object_size_field))
+            if object_size is not None and object_size >= 0:
+                attributes[SPANDATA.AWS_S3_OBJECT_SIZE] = object_size
+
+        if (
+            operation_name == "HeadObject"
+            and "Range" not in ctx.params
+            and "PartNumber" not in ctx.params
+        ):
+            # an un-ranged `HEAD` has no body, so `ContentLength` is the object size.
+            object_size = _as_integer(response.get("ContentLength"))
+            if object_size is not None and object_size >= 0:
+                attributes[SPANDATA.AWS_S3_OBJECT_SIZE] = object_size


Non-ranged GetObject never sets aws.s3.object_size

Un-ranged GetObject only sets http.response.body.size from ContentLength; mirror the HeadObject guard so full-object gets also set aws.s3.object_size.

Evidence

For GetObject, ContentLength is only written to SPANDATA.HTTP_RESPONSE_BODY_SIZE.

aws.s3.object_size from ContentLength is set only for un-ranged HeadObject (Range/PartNumber absent).

GetObject is not in _RESPONSE_OBJECT_SIZE_FIELDS and has no equivalent un-ranged branch.

PR intent: object size from ContentLength on GetObject and HeadObject responses when not a range.

_{Identified by Warden · find-bugs · 6BP-LMY}

Fix attempt detected (commit 85a8895)

The S3 response-attribute code adds an un-ranged HeadObject object-size branch, but still only records ContentLength for GetObject as the HTTP body size, so the reported issue persists.

The original issue appears unresolved. Please review and try again.

_{Evaluated by Warden}

- otherwise service-specific spans with different origins would be skipped (e.g. DynamoDB)

ref(boto3): improve docstring for `get_request_attributes()` Co-authored-by: Erica Pisani <hey@ericapisani.dev>

…t_span_origin()`

pabloDeputter changed the base branch from master to pablo/add-boto3-service-extension-class September 22, 2026 14:50

pabloDeputter added this pull request to stack #7541 September 22, 2026 14:52

pabloDeputter force-pushed the pablo/add-boto3-service-extension-class branch from d8279cc to d7727cd Compare September 22, 2026 14:53

pabloDeputter force-pushed the pablo/add-boto3-s3-service branch from ca45d46 to 0c2873f Compare September 22, 2026 14:53

pabloDeputter removed this pull request from stack #7541 September 22, 2026 14:57

pabloDeputter added this pull request to stack #7618 September 22, 2026 14:57

pabloDeputter force-pushed the pablo/add-boto3-service-extension-class branch from d7727cd to c1249cb Compare September 29, 2026 08:34

pabloDeputter force-pushed the pablo/add-boto3-s3-service branch from 0c2873f to 70218f1 Compare September 29, 2026 08:34

sentry-warden Bot reviewed Sep 29, 2026

View reviewed changes

pabloDeputter force-pushed the pablo/add-boto3-service-extension-class branch from 5c6b2b0 to 5e335b5 Compare September 30, 2026 09:21

pabloDeputter force-pushed the pablo/add-boto3-s3-service branch 2 times, most recently from 2fdaa4b to 85a8895 Compare September 30, 2026 12:04

pabloDeputter force-pushed the pablo/add-boto3-service-extension-class branch from 5e335b5 to 70414fe Compare September 30, 2026 12:04

pabloDeputter removed this pull request from stack #7618 September 30, 2026 12:05

pabloDeputter force-pushed the pablo/add-boto3-s3-service branch from 85a8895 to a0c0e25 Compare September 30, 2026 13:31

pabloDeputter force-pushed the pablo/add-boto3-service-extension-class branch from 70414fe to 0a53853 Compare September 30, 2026 13:31

pabloDeputter added 12 commits September 30, 2026 16:04

ref(boto3): split integration into internal modules

4d1d3f5

ref(boto3): Move consts to

0c23830

fix(boto3): Move imports

b45f76c

ref(boto3): Remove dupliocate _check_minimum_version()

ab843bc

ref(boto3): Update integration checks to use IDENTIFIER constant

b21b888

fix(boto3): harden StreamingBody span finalization

f91d311

lint

ec2a236

ref(boto3): renaming vars

3b53fa4

ref(boto3): Add specific comment on why we initialize with

eea1fc1

fix(boto3): ensure span finishes correctly when reading streaming body

364f547

fix(boto3): trace the complete client-call lifecycle

4f3cfce

fix merging issues

644aff8

pabloDeputter and others added 26 commits September 30, 2026 16:04

add base ServiceExtension class

02a8fa1

feat(boto3): Add service registry

ee7e768

add missing __init__.py file

4d9ae30

simplify _resolve_service() logic

e90b407

add service extension logic to _instrumentation.py

af519a0

combine span-config extraction and validation in _start_client_span()

98cc2a4

add service extension logic to _client.py

ca7e449

add tests

f95f4fe

fix(boto3): Remove "hardcoded" origin

b0b9f0c

- otherwise service-specific spans with different origins would be skipped (e.g. DynamoDB)

feat(boto3): Add sentry.kind attribute to boto span

7d645b4

ref(boto3): Add comment explaining

64e1464

Update sentry_sdk/integrations/boto3/_services/base.py

f221d72

ref(boto3): improve docstring for `get_request_attributes()` Co-authored-by: Erica Pisani <hey@ericapisani.dev>

fix(boto3): correct AWS_RPC_SYSTEM_NAME constant

a106f76

ref(boto3): refactor get_span_config() into get_span_op() and `ge…

032eff3

…t_span_origin()`

ref(boto3): remove attribute-precedence

b1d9551

ref(boto3): remove lazy-loading approach; replace with static map

c2f4f6b

fix(boto3): make logic clearer

a928e74

ref(boto3): change comment's and simplify logic

2d85548

ref(boto3): remove defensive checks

9d2bd96

ref(boto3): remove defensive checks and cleanup comments

b65c0ca

cleanup tests

1270b49

remove AI comment

91c9bdb

feat(attributes): Add s3 relevant attributes to SPANDATA

f2a2bc6

feat(boto3): Register s3 to _SERVICE_EXTENSIONS

0e948cb

feat(boto3): Add attribute extraction helpers

a3629db

feat(boto3): Add S3 extensio

7c39a6c

pabloDeputter force-pushed the pablo/add-boto3-s3-service branch from a0c0e25 to 7c39a6c Compare September 30, 2026 14:05

pabloDeputter force-pushed the pablo/add-boto3-service-extension-class branch 3 times, most recently from 226f56a to 4d0bd32 Compare September 30, 2026 15:50

Navigation Menu

feat(boto3): Add S3 service extension - #7617

feat(boto3): Add S3 service extension#7617
pabloDeputter wants to merge 72 commits into
pablo/add-boto3-service-extension-classfrom
pablo/add-boto3-s3-service

pabloDeputter commented Sep 22, 2026

Uh oh!

github-actions Bot commented Sep 22, 2026 •

edited

Loading

Uh oh!

sentry-warden Bot Sep 29, 2026 •

edited

Loading

Uh oh!

sentry-warden Bot Sep 29, 2026

Uh oh!

sentry-warden Bot Sep 29, 2026 •

edited

Loading

Uh oh!

sentry-warden Bot Sep 29, 2026

Uh oh!

sentry-warden Bot Sep 30, 2026

Uh oh!

This branch has not been deployed

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

pabloDeputter commented Sep 22, 2026

Description

Issues

Uh oh!

github-actions Bot commented Sep 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Results 📊

📊 Comparison with Base Branch

➕ New Tests (1)

❌ Failed Tests

test_cache_spans_item_size[True]

Uh oh!

sentry-warden Bot Sep 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

sentry-warden Bot Sep 29, 2026

Choose a reason for hiding this comment

Uh oh!

sentry-warden Bot Sep 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

sentry-warden Bot Sep 29, 2026

Choose a reason for hiding this comment

Uh oh!

sentry-warden Bot Sep 30, 2026

Choose a reason for hiding this comment

Uh oh!

This branch has not been deployed

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

github-actions Bot commented Sep 22, 2026 •

edited

Loading

`test_cache_spans_item_size[True]`

sentry-warden Bot Sep 29, 2026 •

edited

Loading

sentry-warden Bot Sep 29, 2026 •

edited

Loading