feat: Add FallbackChatGenerator

[vblagoje]

Why

Introduces a resilient chat-generator wrapper so production pipelines can survive provider hiccups while keeping latency predictable and responses flowing.

What

• Added FallbackChatGenerator component that cascades across multiple chat generators with detailed success/failure metadata.
• Registered the component in chat.__init__ and documented behavior with docstrings.
• Created FallbackChatGenerator unit tests covering sync/async execution and error propagation.
• Reno note highlighting why the fallback matters for reliability-focused deployments.

How can it be used

from haystack.components.generators.chat import FallbackChatGenerator, OpenAIChatGenerator

primary = OpenAIChatGenerator(model="gpt-5")
backup = OpenAIChatGenerator(model="gpt-4o-mini")

generator = FallbackChatGenerator(generators=[primary, backup])
result = generator.run(messages=[ChatMessage.from_user("Plan the rollout")])
print(result["meta"]["successful_generator_class"])  # e.g. OpenAIChatGenerator

How did you test it

• pytest test/components/generators/chat/test_fallback.py
• Numerous manual examples and experiments

Notes for the reviewer

Focus on the fallback loop and metadata contract (failed_generators, successful_generator_*); downstream consumers might rely on those keys for telemetry.

[coveralls]

Pull Request Test Coverage Report for Build 18592817487

Warning: This coverage report may be inaccurate.

This pull request's base commit is no longer the HEAD commit of its target branch. This means it includes changes from outside the original pull request, including, potentially, unrelated coverage changes.

• For more information on this, see Tracking coverage changes with pull request builds.
• To avoid this issue with future PRs, see these Recommended CI Configurations.
• For a quick fix, rebase this PR at GitHub. Your next report should be accurate.

Details

• 0 of 0 changed or added relevant lines in 0 files are covered.
• 57 unchanged lines in 11 files lost coverage.
• Overall coverage increased (+0.1%) to 92.2%

---

Files with Coverage Reduction	New Missed Lines	%
components/embedders/sentence_transformers_text_embedder.py	1	96.43%
core/pipeline/pipeline.py	1	92.44%
components/converters/csv.py	4	95.45%
components/generators/azure.py	4	91.84%
components/generators/chat/azure.py	4	93.55%
components/generators/openai.py	4	94.2%
components/agents/agent.py	5	96.77%
core/pipeline/breakpoint.py	5	89.89%
components/generators/chat/openai.py	6	96.76%
core/errors.py	7	89.06%

Totals
Change from base Build 18340248289:	0.1%
Covered Lines:	13346
Relevant Lines:	14475

---

💛 - Coveralls

[sjrl]

for consistency with our other components that accept a ChatGenerator let's call this chat_generators

Suggested change

	def __init__(self, generators: list[ChatGenerator]):
	def __init__(self, chat_generators: list[ChatGenerator]):

[sjrl]

@vblagoje looking very clean! I just had a few comments, I'll let @Amnah199 finish the review :)

[vblagoje]

Thank you @sjrl I'll fix these now!

[Amnah199]

Looks very good overall. Some minor comments, plus maybe we need to update pydocs, generators_api.yml to add the new chat generator.

[Amnah199]

Lets explain these parameters and return instance in docstrings.

[Amnah199]

@vblagoje lets just add consistent docstrings for run methods etc. Otherwise looks good to merge!

[Amnah199]

Looks good. Only docstrings of to_dict and from_dict can be proved. Buts its your call.