chore(deps): update minor and patch

[renovate]

ℹ️ Note

This PR body was truncated due to platform limits.

This PR contains the following updates:

Package	Type	Update	Change	Age	Adoption	Passing	Confidence
alpine (source)	final	minor	3.22 → 3.24
alpine (source)	final	minor	3.22.2 → 3.24.2
alpine (source)	stage	minor	3.22 → 3.24
astroid		minor	==4.0.2 → ==4.3.3
botocore		minor	==1.41.4 → ==1.43.109
docker (source)	stage	minor	29.6.1 → 29.8.2
docker (source)	stage	minor	29.0.4 → 29.8.2
docutils (changelog)		minor	==0.22.3 → ==0.23
golang (source)	final	minor	1.25-alpine3.22 → 1.26-alpine3.22
golang (source)	final	minor	1.25.4-alpine → 1.27.1-alpine
golang (source)	stage	minor	1.25-alpine3.22 → 1.26-alpine3.22
golangci/golangci-lint	stage	minor	v2.6.2-alpine → v2.14.0-alpine
grafana/mimirtool (source)	stage	minor	3.0.0 → 3.2.2
hashicorp/packer	stage	minor	1.15.4 → 1.16.1
hashicorp/packer	stage	minor	1.14.3 → 1.16.1
hashicorp/terraform (source)	stage	minor	1.15.8 → 1.16.5
hashicorp/terraform (source)	stage	minor	1.14.0 → 1.16.5
hashicorp/vault	stage	patch	1.21.1 → 1.21.4
jmespath		minor	==1.0.1 → ==1.1.0
mikefarah/yq (source)	stage	minor	4.53.3 → 4.54.1
mikefarah/yq (source)	stage	minor	4.49.2 → 4.54.1
node (source)	final	minor	25.2-alpine3.22 → 25.9-alpine3.22
node (source)	stage	minor	25.2-alpine3.22 → 25.9-alpine3.22
openpolicyagent/conftest	stage	minor	v0.68.2 → v0.71.1
prom/alertmanager	stage	minor	v0.33.1 → v0.34.1
prom/alertmanager	stage	minor	v0.29.0 → v0.34.1
prom/prometheus	stage	minor	v3.13.1 → v3.15.0
prom/prometheus	stage	minor	v3.7.3 → v3.15.0
prometheus-client		minor	==0.23.1 → ==0.26.0
pylint (changelog)		minor	==4.0.3 → ==4.1.2
s3transfer		minor	==0.15.0 → ==0.19.2
wrapt (changelog)		minor	==2.0.1 → ==2.5.0
yamllint		minor	==1.37.1 → ==1.38.0

---

Release Notes

pylint-dev/astroid (astroid)

v4.3.3

Compare Source

What's new in astroid 4.3.3?

Release date: 2026-09-30

Bug Fixes

• Preserve the length of inferred list and tuple concatenations when an element
  has multiple possible values. An ambiguous or uninferable element is now
  represented as unknown instead of being expanded into multiple sequence items.

  Refs #​3221
  Closes pylint-dev/pylint#2621

• Fix TypeError in ClassDef.getitem when __class_getitem__ is uninferable.

  Closes #​3312

• Fix AssertionError in NodeNG.root() when inferring Class.__bases__, which previously returned a parentless Tuple.

  Closes pylint-dev/pylint#11491

v4.3.2

Compare Source

What's new in astroid 4.3.2?

Release date: 2026-09-25

Bug Fixes

• Fix a crash when a namedtuple or Enum type name or field name
  contains str.format markup, as in namedtuple("{0}", "abc"). The name
  is interpolated into an error message that astroid then reformats, so
  building the message raised IndexError. Inference now falls back to its
  default.

  Closes #​3199

• Infer a functional TypeVar or NewType whose name is not an identifier
  without splicing that name into the synthesized class source. A crafted value
  such as TypeVar("T(Base): #") no longer breaks out of the identifier
  position to inject bases or a body; the name is assigned to the class verbatim.

  Refs #​3217

• Raise InferenceError instead of crashing with IndexError when a class
  uses six.with_metaclass() with no positional metaclass argument, both while
  building the module and when the call is inferred.

  Refs #​3248

• ClassDef.slots() no longer crashes on a class that declares a PEP 695
  type parameter literally named __slots__ (e.g. class C[__slots__]:).
  The type parameter's TypeVar node was mistaken for the class's actual
  __slots__ value; slots() now returns None for such classes
  instead of raising AttributeError.

  Refs #​3258

• Fix a crash when inferring an f-string or str.format() call whose c format
  spec gets an out-of-range code point, such as f"{-1:c}". These now infer to
  Uninferable, like the %c case fixed in #​3254.

  Closes #​3301

• The args attribute of an exception instance is now inferred as a tuple
  instance of unknown contents instead of a known empty tuple, so unpacking
  exc.args no longer looks unbalanced.

  Closes pylint-dev/pylint#11312

v4.3.1

Compare Source

What's New in astroid 4.3.1?

Release date: 2026-08-17

• Fix inference of the attributes of a namedtuple created with
  rename=True. The renamed fields were applied to _fields and to the
  class body, but the instance kept the field names as written, so an instance
  of namedtuple("Tuple", "abc def", rename=True) was inferred as having a
  def attribute instead of _1, and duplicate field names collapsed into
  a single attribute instead of being renamed.

  Closes #​242

v4.3.0

Compare Source

What's New in astroid 4.3.0?

Release date: 2026-08-07

Version 4.2.0 was skipped: a v4.2.0 tag was created by mistake during an
aborted release attempt, so the changes from the 4.2.0 betas are released
as 4.3.0.

• Fix a crash on a functional namedtuple whose field name changes under NFKC
  normalization, such as namedtuple("mu", ["µ"]). Python normalizes
  identifiers, so the parsed class stores the field under "μ" while the
  brain looked it up as written, raising KeyError on a definition that
  namedtuple itself accepts.

  Closes pylint-dev/pylint#8746

• Fix inference of a method's first argument (self or cls) when the
  method's return value is stored on an unrelated object. This avoids inferring
  cls as the caller's class when indexing a tuple returned by a
  classmethod such as A.c()[0].

  Closes pylint-dev/pylint#11101

• super() with no argument now resolves against the class of the object the
  method was called on, instead of the class the method is written in. One consequence
  is that a typing.Self return value survives a chain of super() calls.

  Closes #​2852
  Closes pylint-dev/pylint#10807

• typing.overload stubs no longer shadow the implementation when a special
  method is looked up in the MRO.

  Closes #​2448

• Fix a crash when the field names of a functional Enum or namedtuple
  call are bytes, as in Enum("", b""). Such a definition is invalid, so
  inference now falls back to its default instead of raising TypeError.

  Closes #​3189

• Fix a crash when the body of a typing.NamedTuple subclass contains an
  assignment whose target is not a single name, such as cat.color = "black",
  basket[0] = "apple" or apple, banana = "red", "yellow". The names bound
  by unpacking are now copied to the inferred class as well.

  Closes #​3190

• Fix a crash when a binary operation involves a class whose metaclass is not a
  class, as in class C(metaclass=sum) followed by C | None. The type of
  such a class is a function, which has no bases, so has_known_bases() now
  returns False for anything that is not a class.

  Closes #​3191

• Fix a crash when a dataclass inherits from a dataclass that annotates
  __init__ as a field, as in __init__: int. Such a base binds
  __init__ to a name instead of a function, so it no longer contributes
  arguments to the generated __init__.

  Closes #​3200

• The documentation build now fails on any Sphinx warning, so a cross-reference
  that does not resolve is caught by CI instead of quietly rendering as plain
  text. Uninferable, Position and the BadOperationMessage classes
  are documented as part of the public API, and a number of docstrings that
  Sphinx could not parse were repaired.

• The >>> examples in the documentation are run when the documentation is
  built, so one that stops being true is noticed. Most of them could not run
  before: they printed node addresses recorded in 2018, and the multi-line ones
  were missing the ... continuation prompt. The example for Decorators
  reported the wrong line numbers.

• Fix the example in the docstring of BoolOp, which showed the node it was
  copied from: astroid.extract_node("a and b") gives a BoolOp, not a
  BinOp.

• Names written between single backticks in the ChangeLog render as code again,
  rather than as italics.

• Prevent crashes while processing extension-module classes whose bases cannot
  be resolved when applying enum inference transformations.

  Closes pylint-dev/pylint#11179

• The API documentation now covers the proxies, the objects with no node of
  their own, the inference context, the manager, the object models and the
  extension call signatures. The EvaluatedObject, Interpolation,
  NamedExpr and TemplateStr nodes were missing from the list of nodes
  and are listed now.

• Add __required_keys__ and __optional_keys__ to the inferred
  TypedDict base class, so subclasses no longer raise a no-member
  false positive in pylint when accessing those attributes at runtime.

  Closes pylint-dev/pylint#10158

• Comprehension conditions now constrain the inference of names used in the
  comprehension, like if statement conditions already did:
  [x for x in lst if x is not None] no longer infers None for x
  in the element expression.

  Refs #​3094

• Fix isinstance() constraints only filtering the first inferred value of
  a variable: checking a value made the isinstance classinfo uninferable
  for the checks of the following values.

• Remove Python 2 era dead code and outdated comments.
  Notably: the internal TreeRebuilder constructor no longer takes a
  parser_module argument.

  Refs #​3154

• Fix inference of a starred target that is not last in a for loop, such as
  for a, *b, c in .... The elements following the starred one were counted
  from the target's own arity rather than from the end of the iterable, so
  b was truncated whenever the iterable was longer than the target.

• Add qname() method to the Slice node class, returning
  "builtins.slice", so that the node can be used in contexts
  that require a qualified name (e.g. pylint's basic_checker).

  Closes #​3115

• Add brain tip for numpy.fromfile so that calls to
  np.fromfile(...) are correctly inferred as returning
  numpy.ndarray instead of Uninferable.

  Closes #​600

• Add brain for decimal that gets applied if _decimal isn't available.

• Removed the **kwargs argument from all infer functions.
  Users upgrading their astroid versions should ensure they do no pass values other than
  context to infer.

• Deduplicate keys when inferring dict.fromkeys. dict.fromkeys("aab")
  now infers a dict with keys "a", "b" instead of "a", "a",
  "b", matching the runtime. This also stops dict.fromkeys from
  materializing one Const node per character for a repeated string such as
  dict.fromkeys("x" * 10 ** 8), which is a single-key dict.

• Bound str/bytes and list/tuple concatenation (a + b) during inference.

• Bound oversized old-style (%) string/bytes formatting during inference.
  A tiny literal such as "%1000000000d" % 1 made
  _infer_old_style_string_formatting eagerly build a multi-gigabyte
  Const while inferring otherwise untrusted source. The width and
  precision are now read out of the conversion specifiers (including *
  fields) and the interpolation infers as Uninferable past 1e8,
  mirroring the repetition, concatenation and str.format caps. Bytes
  % formatting is routed through the same handler so it is bounded too;
  small format strings keep inferring their exact value.

• Fix crash (TypeError: 'UninferableBase' object is not iterable) in the
  multiprocessing brain when a local package shadows the stdlib
  multiprocessing module.

  Closes pylint-dev/pylint#10014

• Bound the number of nodes built when inferring list/set/tuple/
  frozenset and dict.fromkeys from a str/bytes constant. These
  built one Const node per character with no cap, so ``list(("a" * 10 ** 8)

  • "b")(concatenation is not size-bounded) materialized hundreds of millions of nodes. They now fall back to the default inference past1e8`` characters,
    matching the sequence-repetition guard.

Refs #​3127

• Bound string/bytes multiplication ("x" * n) and integer left shifts
  (1 << n) in const_infer_binary_op the same way list/tuple
  multiplication and ** already are. Inferring a constant such as
  "A" * 10 ** 10 previously materialized the multi-gigabyte result
  eagerly; these operations now infer as Uninferable when the result
  would be oversized.

  Refs #​3107

• Bound the field width and precision when inferring str.format calls.
  A template such as "{:>2000000000}".format("x") (or a width/precision
  supplied through a nested {} field) previously built the padded
  multi-gigabyte string eagerly during inference; such calls now infer as
  Uninferable when the field size would be oversized.

  Refs #​3131

• Remove the asname keyword argument from Import._infer and
  ImportFrom._infer. The alias-to-real-name translation that
  asname=True performed is now hoisted into ImportNode._infer_name,
  which runs in _infer_stmts before _infer is dispatched. Direct
  callers of Import.infer / ImportFrom.infer (previously the
  asname=False path) now consistently resolve the lookup name as-is.

  The two flows used to collapse into a single inference cache entry
  because the asname kwarg was not part of the cache key, so the
  second call returned the cached result of the first regardless of
  which one ran. With the translation hoisted upstream, the two flows
  set different lookupname values and therefore land on distinct
  cache keys, eliminating the collision.

  Refs pylint-dev/pylint#10193
  Closes #​3007

• Avoid materializing a multi-gigabyte string while inferring a small literal
  such as "{:>2000000000}".format("x") or f"{1.5:.2000000000f}".
  _infer_str_format_call and FormattedValue._infer now yield
  Uninferable when a format spec asks for a width or precision over 1e8,
  mirroring the sequence/repetition caps in astroid.protocols.

Refs #​3108

• Fix uncaught IndentationError when parsing code whose lines end in
  \r: the slice of source tokenized to compute a class or function
  position is split on \n only and could be misaligned with the AST
  line numbers. Such nodes now simply have no position information, as
  already done when tokenize raises TokenError.

  Closes #​3091

• Support PEP 810 lazy imports (new in Python 3.15). Import and
  ImportFrom nodes gain an is_lazy integer attribute, mirroring
  the ast field of the same name.

• Support PEP 798 comprehension unpacking ({**d for d in dicts},
  new in Python 3.15).

• Fix astroid bootstrap crash on PyPy 7.3.22 (TypeError: expected str, got getset_descriptor object) by also catching TypeError from
  getattr(obj, alias) in InspectBuilder.object_build. PyPy 7.3.22
  raises TypeError instead of AttributeError for unset getset
  descriptors like types.FunctionType.__text_signature__, which made
  _astroid_bootstrapping() blow up on any call into astroid.

  Refs pylint-dev/pylint#10999

• Shorten import astroid by deferring imports only needed on cold paths.
  logging (used by modutils and raw_building solely to report
  stderr/stdout captured while importing a module) and pprint (used only
  to format debug __repr__ / repr_tree output) are now imported
  lazily, and astroid.exceptions imports astroid.typing under
  TYPE_CHECKING.

• Fix AttributeError crash in starred_assigned_stmts when a starred
  unpacking target is an attribute (e.g. for *o.attr, x in ...) rather
  than a simple name.

  Closes #​2646

• Fix AttributeError crash in the Arguments assigned_stmts
  protocol when called without an inference context (the public
  assigned_stmts API defaults context to None). Resolving a
  function's first parameter dereferenced context.boundnode while the
  matching context and ... guard a few lines below was missing; it now
  degrades to Uninferable.

• Fix AttributeError crash in ClassDef.infer_call_result when called
  through the public API without a context (it defaults to None) for a
  class whose metaclass defines __call__: the callee was assigned to
  context.callcontext.callee without checking that a call context exists.
  It is now guarded, matching the surrounding code.

• Wrap assignment expressions (:=) in parentheses when emitting
  as_string output so the rendered code remains syntactically valid in
  contexts such as comparisons, where Python requires the walrus expression
  to be parenthesized.

  Closes #​2668

• Catch MemoryError/RecursionError (and ValueError) when validating
  type comments with ast.parse. Pathological type comments produced by
  fuzzers (e.g. # type: i{{{{{{{...) previously crashed parsing with a
  MemoryError or RecursionError depending on the runtime; astroid now
  treats them as invalid type comments and skips them, mirroring the f-string
  fix from #​2762.

  Closes #​2993

• Bypass __init__ in InferenceContext.clone() and write the slots
  directly. clone() is called ~85k times per pandas/frame.py pylint run;
  skipping the conditional defaults shaves measurable time off the hottest
  constructor in inference.

  Refs #​1115

• Fix TypeError in brain_random when random.sample is called with a
  sequence containing nodes whose __init__ does not accept lineno
  (e.g. Module). The clone helper now filters init params to those the
  class actually accepts.

  Closes #​3043

• Fix RecursionError in _compute_mro() when circular class hierarchies
  are created through runtime name rebinding. Circular bases are now resolved
  to the original class instead of recursing.

  Closes #​3023
  Closes pylint-dev/pylint#10821

• Changed block_range to consider else its own block, allowing pylint to apply
  disables to just the block.

  References pylint-dev/pylint#872

• Fix uncaught TokenError when building a class or function whose source
  slice is malformed. tokenize.generate_tokens may raise (e.g. on an
  unterminated bracket on Python < 3.12); position computation now treats such
  a node as having no position information instead of crashing.

  Closes #​2527

• str() of a constant argument now infers the actual string value instead
  of always inferring "". When every inference path of the argument
  resolves to Const values that stringify to the same string, infer_str
  returns that string; otherwise it keeps falling back to Const("").

  Closes #​2994

• Fix AttributeError crash when looking up a special method on a class
  whose explicit metaclass infers to a non-class node (e.g. a function). Such
  a metaclass has no MRO, so the dunder lookup now raises
  AttributeInferenceError instead of crashing.

  Closes #​3063

• Fix AttributeError crash in ClassDef.getitem when __class_getitem__
  resolves to a non-callable node (e.g. an AssignName). getitem now
  raises AstroidTypeError in that case, consistent with its documented
  behaviour.

  Closes #​3064

• Fix DuplicateBasesError crash in the enum brain when inferring an enum
  class with duplicate bases (e.g. class C(enum.Enum, enum.Enum)).
  infer_enum_class now catches MroError and leaves such a malformed
  class untransformed.

  Closes #​3065

• Fix InferenceError crash in ClassDef.slots() when __slots__ is
  declared as an annotation without a value (e.g. __slots__: None). Such a
  __slots__ cannot be inferred, so slots() now returns None.

  Closes #​3067

• Fix TypeError crash in the enum brain when a functional Enum call
  has a non-string member name (e.g. Enum("e", (1,))). Such a definition
  is invalid, so inference now falls back to the default instead of crashing.

  Closes #​3068

• Fix detecting static/class methods and inspecting IntFlag types in
  GObject-based libraries (GLib, Gtk etc)

• Only inject the _HAS_DEFAULT_FACTORY sentinel into a module's locals when
  the generated dataclass __init__ actually references it. Parsing a module
  containing a dataclass without any field(default_factory=...) no longer
  exposes an unexpected _HAS_DEFAULT_FACTORY name.

  Closes #​2808

  • Fix a crash when inferring __func__ on a bound method that proxies its function directly,
    such as A.method.__func__ for a classmethod, or a lambda assigned to a class attribute
    such as A().lam.__func__.

  Closes pylint-dev/pylint#11198

• Add qname() and pytype() to the TypeVar, ParamSpec,
  TypeVarTuple and TypeAlias node classes, returning
  "typing.TypeVar", "typing.ParamSpec", "typing.TypeVarTuple"
  and "typing.TypeAliasType". PEP 695 type parameters and type aliases
  are inferred as themselves, so callers that ask an inferred value for its
  type crashed with an AttributeError, as FunctionDef.decoratornames()
  did for @T inside class Basket[T].

  Refs #​3115

• Infer the value of a data descriptor of a C-implemented class as unknown
  instead of as a class named after the attribute. Reading such an attribute,
  for instance exc.__traceback__.tb_frame or gen.gi_frame, inferred a
  class, so the attributes of the value the descriptor returns were reported as
  missing. raw_building.object_build_datadescriptor() now returns an
  EmptyNode.

  Closes pylint-dev/pylint#11218

v4.1.2

Compare Source

============================
Release date: 2026-03-22

• Fix crash accessing property fset in generic classes with type annotations.
  Closes #​2996

• Fix infinite recursion caused by cyclic inference in Constraint.

• Fix RecursionError in _compute_mro() when circular class hierarchies
  are created through runtime name rebinding. Circular bases are now resolved
  to the original class instead of recursing.

  Closes #​2967
  Closes pylint-dev/pylint#10821

• Fix DuplicateBasesError crash in dataclass transform when a class has
  duplicate bases in its MRO (e.g., Protocol appearing both directly and
  indirectly). Catch MroError at .mro() call sites in
  brain_dataclasses.py, consistent with the existing pattern elsewhere.

  Closes #​2628

• Fix FunctionModel returning descriptor attributes for builtin functions.

  Closes #​2743

• Catch MemoryError when inferring f-strings with extremely large format
  widths (e.g. f'{0:11111111111}') so that inference yields Uninferable
  instead of crashing.

  Closes #​2762

• Fix ValueError in __str__/repr and error messages when nodes have
  extreme values (very long identifiers or large integers). Clamp pprint width
  to a minimum of 1 and truncate oversized values in error messages.

  Closes #​2764

v4.1.1

Compare Source

============================
Release date: 2026-02-22

• Let UnboundMethodModel inherit from FunctionModel to improve inference of
  dunder methods for unbound methods.

  Refs #​2741

• Filter Unknown from UnboundMethod and Super special attribute
  lookup to prevent placeholder nodes from leaking during inference.

  Refs #​2741

v4.1.0

Compare Source

============================
Release date: 2026-02-08

• Add support for equality constraints (==, !=) in inference.
  Closes pylint-dev/pylint#3632
  Closes pylint-dev/pylint#3633

• Ensure ast.JoinedStr nodes are Uninferable when the ast.FormattedValue is
  Uninferable. This prevents unexpected-keyword-arg messages in Pylint
  where the Uninferable string appeared in function arguments that were
  constructed dynamically.

  Closes pylint-dev/pylint#10822

• Add support for type constraints (isinstance(x, y)) in inference.

  Closes pylint-dev/pylint#1162
  Closes pylint-dev/pylint#4635
  Closes pylint-dev/pylint#10469

• Make type.__new__() raise clear errors instead of returning None

• Move object dunder methods from FunctionModel to ObjectModel to make them
  available on all object types, not just functions.

  Closes #​2742
  Closes #​2741
  Closes pylint-dev/pylint#6094

• lineno and end_lineno are now available on Arguments.

• Add helper to iterate over all annotations nodes of function arguments,
  Arguments.get_annotations().

  Refs #​2860

• Skip direct parent when determining the Decorator frame.

  Refs pylint-dev/pylint#8425

• Add simple command line interface for astroid to output generated AST.
  Use with python -m astroid.

• Fix incorrect type inference for super().method() calls that return Self.
  Previously, astroid would infer the parent class type instead of the child class type,
  causing pylint E1101 false positives in method chaining scenarios.

  Closes #​457

• Add missing dtype and casting parameters to numpy.concatenate brain.

  Closes #​2870

• Fix ability to detect .py modules inside PATH directories on Windows
  described by a UNC path with a trailing backslash (\)

  • Example: modutils.modpath_from_file(filename=r"\Mac\Code\tests\test_resources.py", path=["\mac\code"]) == ['tests', 'test_resources']

• Fix random.sample inference crash when sequence contains uninferable elements.

  Closes #​2518

• Fix random.sample crash when cloning ClassDef or FunctionDef nodes.

  Closes #​2923

v4.0.4

Compare Source

============================
Release date: 2026-02-07

• Fix is_namespace() crash when search locations contain pathlib.Path objects.

  Closes #​2942

v4.0.3

Compare Source

============================
Release date: 2026-01-03

• Fix inference of IfExp (ternary expression) nodes to avoid prematurely narrowing
  results in the face of inference ambiguity.

  Closes #​2899

• Fix base class inference for dataclasses using the PEP 695 typing syntax.

  Refs pylint-dev/pylint#10788

boto/botocore (botocore)

v1.43.109

Compare Source

========

• api-change:lambda-web: Removes operations that are not yet generally available from the Lambda Web.

v1.43.108

Compare Source

========

• api-change:cognito-idp: Amazon Cognito User Pools now supports the OIDC-standard authentication context class reference (ACR) and authentication methods reference (AMR) claims on issued access and Id tokens. Amazon Cognito User Pools also now supports step-up authentication via our existing authentication APIs.
• api-change:glue: Added refresh token grant type to Glue Connection supported OAuth 2.0 grant types
• api-change:invoicing: API and doc updates related to adding MarketplacePunchOutEnabled and MarketplacePunchOutPreference fields to ProcurementPortalPreferences related APIs
• api-change:lambda-web: Documentation update for AWS Lambda Web Functions, clarifies that the LambdaWeb APIs are experimental and not yet available to external customers.
• api-change:mediapackagev2: Dynamic Multiview enables viewers to watch multiple live video streams in a single combined output. Static filter configuration allows users to configure endpoints with layouts and sources without using query parameters. The number of sources per multiview channel has been increased to 50.
• api-change:pinpoint-sms-voice-v2: AWS End User Messaging SMS CarrierLookup API now supports phone number cleansing on customer opt-in. when selected, the response includes the additional field "OriginalPhoneNumber". It can also return additional PhoneNumberType enums, VOIP and PREPAID.
• api-change:securityagent: Adds trigger filters that control which pull request events, target branches, and labels start an automatic code review.

v1.43.107

Compare Source

========

• api-change:bedrock-agent: Adds an optional textReadyAt field to ListIngestionJobs and GetIngestionJob for Managed Knowledge Bases data source sync jobs. The field denotes the timestamp at which all the documents in the scope of a sync job had their text content indexed and are available for retrieval.
• api-change:cloudfront: Added always-amz-auth as a supported signing behavior for Origin Access Control (OAC), enabling CloudFront to authenticate requests to Lambda-Web origins.
• api-change:ec2: This release launches the AMI tag sharing feature, which lets AMI owners share tags alongside their AMIs, eliminating the need to build and maintain custom tag replication workflows.
• api-change:endpoint-rules: Update endpoint-rules client to latest version
• api-change:endusermessaging: AWS End User Messaging now supports Brand profiles and Notify code configurations. Brand profiles capture your sender details once to reuse across phone number registrations. Notify code configurations let you define your OTP policy and delivery settings to send passcodes in minutes.
• api-change:health: Adds DescribeServiceLifecycle operation returning lifecycle information for AWS services, including end-of-support dates, version recommendations, and lifecycle events.
• api-change:lambda-web: Lambda Web Functions GA launch. Lambda Web Functions enable customers to run web applications and API backends
• api-change:quicksight: This release adds HierarchyFilter support for Amazon QuickSight analysis and dashboard and 2 legged OAuth for databricks datasources.
• api-change:sagemaker: Release support for c8a.16xlarge and m8a.16xlarge instance types for SageMaker HyperPod
• api-change:securityhub: Adds GetRemediationsV2 and ListExposuresByRemediationV2 APIs. This feature allows customers to see their highest priority remediations for their Exposure findings. Remediations target key changes customers can make to resources to drive finding resolution.
• api-change:transfer: AWS Transfer Family Workflows adds support for the structuredLogDestinations option, enabling customers to specify a custom Amazon CloudWatch Logs log group for managed workflow execution logs.

v1.43.106

Compare Source

========

• api-change:account: This release adds support for verifying an AWS account's primary contact phone number. SendPhoneNumberVerification sends a one-time code by SMS, VerifyPhoneNumber validates it, and GetContactInformation now returns the verification status.
• api-change:agent-registry: Minor doc update for the AWS Agent Registry Custom metadata SearchDiscoverableRegistryRecords API
• api-change:batch: AWS Batch adds support for Amazon EKS access entries on EKS compute environments through the new accessEntry setting in CreateComputeEnvironment and UpdateComputeEnvironment.
• api-change:bedrock: Amazon Bedrock Automated Reasoning policies now accept Unicode letters in identifier names such as type names, type value names, and variable names. You can now author policies in non-English languages using accented or non-Latin characters.
• api-change:bedrock-agentcore-control: This release adds support for private certificate authorities on Amazon Bedrock AgentCore Gateway targets. The new certificateConfigurations parameter on CreateGatewayTarget and UpdateGatewayTarget references a PEM-encoded CA certificate in Amazon S3 or AWS Secrets Manager.
• api-change:connect: Amazon Connect Rules can now trigger in-app notifications to users as a rule action. Notification character limit was increased to 500 visible characters.
• api-change:datazone: Support for setting notebook run notification configurations
• api-change:dynamodb: Adds support for filtering exported table data using FilterExpression, ProjectionExpression and KeyConditionExpression with ExportTableToPointInTime.
• api-change:ecs: Releasing VPCL for BlueGreen ecs deployments.
• api-change:globalaccelerator: IpSets now include the Network Zone for each Static IP address.
• api-change:glue: Enable Catalog ID for crawler, column statistics and materialized views.
• api-change:guardduty: GuardDuty AWS Organizations policy integration. GetDetector and GetMemberDetectors now show whether a GuardDuty policy manages a feature.
• api-change:logs: Amazon CloudWatch Logs now supports an optional roleArn parameter on PutDeliveryDestination for X-Ray trace delivery destinations, specifying the IAM role to assume when delivering traces.
• api-change:observabilityadmin: Enablement for Bedrock PaymentManager logs via Observability Admin Telemetry Rule
• api-change:organizations: Add support for policy operations on the GUARDDUTY POLICY policy type.
• api-change:s3: Amazon S3 adds a new optional S3 Inventory field, IntelligentTieringReferenceDate, reporting the reference date S3 Intelligent-Tiering uses to evaluate an object's tier-transition eligibility. The value is populated for objects in the Intelligent-Tiering storage class and left blank for others.
• api-change:s3vectors: Amazon S3 Vectors now supports metadata prefiltering, providing higher recall on filtered queries.
• api-change:sagemaker: This feature enables customers to modify their accounting database via API.

v1.43.105

[Compare Source](https://redirect.github.com/boto/botocore/comp

❗ Important

✂ PR body was truncated to here.

---

Configuration

📅 Schedule: (UTC)

• Branch creation
  • Only on Sunday and Saturday (* * * * 0,6)
• Automerge
  • At any time (no schedule defined)

🚦 Automerge: Enabled.

♻ Rebasing: Whenever PR is behind base branch, or you tick the rebase/retry checkbox.

👻 Immortal: This PR will be recreated if closed unmerged. Get config help if that's undesired.

---

• If you want to rebase/retry this PR, check this box

---

This PR was generated by Mend Renovate. View the repository job log.