perf: Reduce overhead of flushing stdout stream for every tap record

[edgarrmondragon]

SSIA

Related

• test: Measure performance of writing Singer stream to stdout #3673

Summary by Sourcery

Enhancements:

• Avoid flushing stdout after each record message in both msgspec and simple encoders, flushing only for non-record messages instead.

[sourcery-ai]

Reviewer's guide (collapsed on small PRs)

Reviewer's Guide

Optimizes stdout flushing in Singer SDK message writers by skipping flushes for high-volume RecordMessage instances while retaining flush behavior for other message types.

Sequence diagram for conditional stdout flushing in write_message

sequenceDiagram
    participant Writer
    participant sys_stdout

    Writer->>sys_stdout: write
    alt [message is RecordMessage]
        Note over Writer,sys_stdout: No flush for high-volume RecordMessage
    else [message is not RecordMessage]
        Writer->>sys_stdout: flush
    end

Loading

File-Level Changes

Change	Details	Files
Avoid flushing stdout on every record message to reduce I/O overhead while preserving flushes for non-record messages.	Extend msgspec encoding module to import RecordMessage type along with Message. Update msgspec-based writer to write bytes to sys.stdout.buffer and only flush when the message is not a RecordMessage. Update simple writer implementation to append a newline, write to sys.stdout, and only flush when the message is not a RecordMessage.	singer_sdk/contrib/msgspec.py singer_sdk/singerlib/encoding/simple.py

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it. You can also reply to a
  review comment with @sourcery-ai issue to create an issue from it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time. You can also comment
  @sourcery-ai title on the pull request to (re-)generate the title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time exactly where you
  want it. You can also comment @sourcery-ai summary on the pull request to
  (re-)generate the summary at any time.
• Generate reviewer's guide: Comment @sourcery-ai guide on the pull
  request to (re-)generate the reviewer's guide at any time.
• Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
  pull request to resolve all Sourcery comments. Useful if you've already
  addressed all the comments and don't want to see them anymore.
• Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
  request to dismiss all existing Sourcery reviews. Especially useful if you
  want to start fresh with a new review - don't forget to comment
  @sourcery-ai review to trigger a new review!

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.

[sourcery-ai]

Hey - I've found 1 issue, and left some high level feedback:

• Since RecordMessage writes are now buffered without an immediate flush, consider adding a clear comment or helper function documenting where/when the final flush is expected to occur to avoid confusion about output ordering and buffering behavior.
• The conditional flush logic is duplicated in both msgspec and simple writers; consider centralizing this behavior in a shared helper or base class method to keep the flush semantics consistent and easier to maintain.

Prompt for AI Agents

Please address the comments from this code review:

## Overall Comments
- Since `RecordMessage` writes are now buffered without an immediate flush, consider adding a clear comment or helper function documenting where/when the final flush is expected to occur to avoid confusion about output ordering and buffering behavior.
- The conditional flush logic is duplicated in both `msgspec` and `simple` writers; consider centralizing this behavior in a shared helper or base class method to keep the flush semantics consistent and easier to maintain.

## Individual Comments

### Comment 1
<location path="singer_sdk/contrib/msgspec.py" line_range="120" />
<code_context>
         """
         sys.stdout.buffer.write(self.format_message(message))
-        sys.stdout.flush()
+        if not isinstance(message, RecordMessage):
+            sys.stdout.flush()
</code_context>
<issue_to_address>
**issue (bug_risk):** Changing flush behavior only for non-record messages may impact streaming semantics for some consumers.

Previously, all messages were flushed immediately; now `RecordMessage` output may sit in the buffer indefinitely unless another (non-record) message is written. For long-running processes or consumers that expect records to appear as soon as they’re written, this can change behavior (e.g., increased latency or seemingly stuck pipelines). If the goal is to batch records for performance, consider adding a configurable flush policy (e.g., count- or time-based) or an explicit API to force flushing so callers can opt into immediate delivery when needed.
</issue_to_address>

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}

[sourcery-ai]

issue (bug_risk): Changing flush behavior only for non-record messages may impact streaming semantics for some consumers.

Previously, all messages were flushed immediately; now RecordMessage output may sit in the buffer indefinitely unless another (non-record) message is written. For long-running processes or consumers that expect records to appear as soon as they’re written, this can change behavior (e.g., increased latency or seemingly stuck pipelines). If the goal is to batch records for performance, consider adding a configurable flush policy (e.g., count- or time-based) or an explicit API to force flushing so callers can opt into immediate delivery when needed.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 94.13%. Comparing base (f2b5a3e) to head (a66352b).

Additional details and impacted files

@@                          Coverage Diff                           @@
##           test/measure-perf-singer-stream-write    #3674   +/-   ##
======================================================================
  Coverage                                  94.12%   94.13%           
======================================================================
  Files                                         73       73           
  Lines                                       6200     6202    +2     
  Branches                                     762      764    +2     
======================================================================
+ Hits                                        5836     5838    +2     
  Misses                                       270      270           
  Partials                                      94       94           

Flag	Coverage Δ
core	82.94% <40.00%> (-0.02%)	⬇️
end-to-end	76.03% <100.00%> (+<0.01%)	⬆️
optional-components	44.79% <20.00%> (-0.04%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[codspeed-hq]

Merging this PR will improve performance by 42.78%

⚡ 1 improved benchmark
✅ 10 untouched benchmarks

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
⚡	test_bench_write_record_messages_msgspec	7.2 ms	5.1 ms	+42.78%

Tip

Curious why this is faster? Comment @codspeedbot explain why this is faster on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing perf/record-flush-overhead (a66352b) with test/measure-perf-singer-stream-write (f2b5a3e)}