Hitless upgrade: Support initial implementation for synchronous Redis client - no handshake, no failing over notifications support.

[petyaslavova]

Pull Request check-list

Please make sure to review and check all of these items:

• Do tests and lints pass with this change?
• Do the CI tests pass with this change (enable it first in your forked repo and wait for the github action build to finish)?
• Is the new or changed code fully tested?
• Is a documentation update included (if this change modifies existing APIs, or introduces new ones)?
• Is there an example added to the examples folder (if applicable)?

NOTE: these things are not required to open a PR and can be done
afterwards / while the PR is open.

Description of change

Hitless upgrade support implementation for synchronous Redis client.

[Copilot]

Pull Request Overview

This PR implements hitless upgrade support for the synchronous Redis client by adding maintenance events handling. The implementation allows the Redis client to gracefully handle cluster rebalancing and node migration operations without losing connections or data.

Key changes:

• Added comprehensive maintenance events handling infrastructure for MOVING, MIGRATING, and MIGRATED events
• Implemented connection pool management for hitless upgrades during cluster maintenance
• Enhanced Redis client with maintenance events configuration and handlers

Reviewed Changes

Copilot reviewed 9 out of 9 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
redis/maintenance_events.py	Core maintenance events classes and handlers for processing cluster rebalancing notifications
redis/connection.py	Enhanced connection and pool classes with maintenance state management and proactive reconnection support
redis/client.py	Updated Redis client to support maintenance events configuration and handle connection reconnection during maintenance
redis/_parsers/base.py	Extended parsers to handle maintenance-related push notifications (MOVING, MIGRATING, MIGRATED)
redis/_parsers/resp3.py	Updated RESP3 parser to properly handle push notifications during maintenance events
redis/_parsers/hiredis.py	Updated Hiredis parser to support maintenance event push notifications
tests/test_maintenance_events_handling.py	Comprehensive integration tests for maintenance events handling with mocked Redis protocol
tests/test_maintenance_events.py	Unit tests for maintenance event classes and handlers
tests/test_connection_pool.py	Minor test updates to support new connection pool functionality

Comments suppressed due to low confidence (2)

redis/connection.py:385

• The parser is being set before protocol-specific configurations are applied. Line 407 shows that RESP3Parser is set when protocol is 3, but the initial parser setup happens before this check. This could lead to incorrect parser being used for maintenance events.

        self.health_check_interval = health_check_interval

redis/connection.py:381

• This line appears to be indented incorrectly. The if statement at line 380 should control this block, but the indentation suggests it's not properly nested within the conditional block.

            # Update the retry's supported errors with the specified errors

[vladvildanov]

I think this flag is redundant. If you try to release a lock that wasn't acquired by this thread it will throw a RuntimeException, if you try to release a lock that was acquired by another thread it will be the same. Since, you anyway surround release() method with try catch, you will either release lock if it's possible or suppress any exception and move on

[petyaslavova]

That's true, but it will be slower.
We will be in the True case very rarely - only when we have active moving maintenance.
If I remove the check, the exception will be raised all the time, and it is a much more resource-consuming operation than the self._locked check.

[vladvildanov]

Why do you need to convert it into a tuple? Set is also iterable

[petyaslavova]

Because I'm changing the elements if any of the existing inside notifications have expired - I can't iterate on an iterable object and change its elements inside the cycle.

[vladvildanov]

If we don't expect to handle any other events then NodeMovingEvent why does signature allows it's parent class to be passed (MaintenanceEvent)

[petyaslavova]

Because we don't know if this will be the only one - it is for now.