fix: update retry strategy for mutation calls to handle aborted transactions

[aakashanandg]

Updating retry strategy for mutation calls to handle aborted transactions

• This PR updates the retry strategy for mutation calls to handle aborted transactions more effectively. Previously, the retry mechanism didn't handle certain edge cases for aborted transactions, leading to failures. The updated strategy ensures retries in these scenarios to improve the robustness of the mutation operations.

Test Results:

• All unit tests related to the retry logic and mutation operations pass successfully.

Fixes #<1133> 🦕

[olavloite]

Hmmm.... Now that I see this, I realize that this cannot use run_in_transaction directly. The reason is that:

1. This method batch.commit() in its original form (so without the changes in this pull request) creates and commits a single-use read/write transaction.
2. run_in_transaction however creates a new read/write transaction and then executes the given method in the scope of that transaction.

So what happens now is that you create two transactions:

1. run_in_transaction creates a transaction that is not being used.
2. method that is passed in to run_in_transaction executes a Commit in a single-use read/write transaction.

[aakashanandg]

Thank you for pointing that out. I've updated the PR with a new retry strategy, replicating the same retry behavior as in the run_in_transaction method.

[olavloite]

You should verify the exact number. Based on my comment above, my guess is that you are getting 1 more commit than you would expect, as run_in_transaction also creates a read/write transaction.

[aakashanandg]

Thank you for the clarification. I’ve added updated assertions to track the number of times api.commit is called with the next revision

[aakashanandg]

A new PR has been created to fix the commit history. Please refer to the new PR for the updated changes. This PR will be closed to keep the commit history clean and in line with the intended changes.
New PR: #1279