fix(bigtable): recycle channel on consecutive new stream failures

[nimf]

No description provided.

[gemini-code-assist]

Code Review

This pull request introduces a mechanism to recycle gRPC channels in the ChannelPoolDpImpl after a threshold of five consecutive session failures. It also implements an exponential backoff for channel recycling to prevent aggressive recreation during outages. Feedback suggests capping the exponential backoff duration to ensure timely recovery and using the subtraction idiom for System.nanoTime() comparisons to correctly handle potential numerical wrap-around.

[gemini-code-assist]

Code Review

This pull request introduces a mechanism to recycle gRPC channels in the Bigtable client after a threshold of consecutive session failures is reached. It includes an exponential backoff strategy for recycling to prevent rapid churn and adds comprehensive unit tests to verify the failure counting and reset logic. Feedback focuses on improving the robustness of the backoff implementation by using the subtraction idiom for System.nanoTime() comparisons and enforcing a maximum bound for the backoff duration to prevent overflow. Suggestions were also made to use constants for backoff durations to improve maintainability.