Fix lossiness when submitting timestamped custom metrics

[nhulston]

What does this PR do?

Fixes an issue in ThreadStatsWriter.flush() where constant_tags were accumulating duplicate tags over multiple Lambda invocations, leading to metric loss for metrics with timestamps. This code was the issue:

if tags:
    self.thread_stats.constant_tags = self.thread_stats.constant_tags + tags

Since self.thread_stats.constant_tags persists across multiple invocations within the same Lambda execution context, we kept appending duplicate tags after each invocation.

This issue only affected metrics with a timestamp, since they use ThreadStatsWriter for submission. After ~12 invocations, the payloads were too large, so custom metrics were being dropped.

This PR modifies the flush() method to still add the provided tags, but it now resets constant_tags to its original state afterward.

Motivation

#514

Testing Guidelines

Manual testing showed that metrics were no longer being dropped after these changes.

Added unit tests

Additional Notes

For people who are more familiar with this library, please make sure that these changes don't have any side effects that I did not consider.

Types of Changes

• Bug fix
• New feature
• Breaking change
• Misc (docs, refactoring, dependency upgrade, etc.)

Check all that apply

• This PR's description is comprehensive
• This PR contains breaking changes that are documented in the description
• This PR introduces new APIs or parameters that are documented and unlikely to change in the foreseeable future
• This PR impacts documentation, and it has been updated (or a ticket has been logged)
• This PR's changes are covered by the automated tests
• This PR collects user input/sensitive content into Datadog
• This PR passes the integration tests (ask a Datadog member to run the tests)

[purple4reina]

good find!

[tianchu]

Great finding!