fix(poll): track consequent errors during polling W-18203875 #1663
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Increase timeout from 1 to 3 seconds in retry limit tests to ensure the retry limit is reached before timeout on all platforms. The 1-second timeout was causing race conditions on Windows where only 16 retries completed instead of the expected 20 due to execution overhead. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Replace total retry limit with consecutive error retry limit to prevent infinite loops from repeated errors while allowing long-running operations to poll indefinitely until timeout. Key changes: - Track consecutive retryable errors separately from normal polling - Reset counter on successful status check - Default limit of 25 consecutive errors (configurable via SF_METADATA_POLL_ERROR_RETRY_LIMIT) - Remove PollingClient retryLimit to allow unlimited normal polling - Add error message for retry limit exceeded without wrapper duplication 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
cristiand391
commented
Jan 6, 2026
src/client/metadataTransfer.ts
Outdated
| this.errorRetryLimit = calculateErrorRetryLimit(this.logger); | ||
| this.errorRetryLimitExceeded = undefined; | ||
|
|
||
| // Set a very high retryLimit for PollingClient to prevent it from stopping on errors |
Member
Author
There was a problem hiding this comment.
todo for me:
remove this comment, claude was still setting a high retryLimit after last changes
cristiand391
commented
Jan 6, 2026
| const err = e as Error | SfError; | ||
|
|
||
| // Don't wrap the error retry limit exceeded error | ||
| if (err instanceof SfError && err.message.includes('consecutive retryable errors')) { |
Member
Author
There was a problem hiding this comment.
avoid wrapping mostly bc the final error printed to the user had duplicate msg parts like Metadata API request failed: Metadata API request failed
cristiand391
commented
Jan 6, 2026
| try { | ||
| mdapiStatus = await this.checkStatus(); | ||
| // Reset error counter on successful status check | ||
| this.consecutiveErrorRetries = 0; |
Member
Author
There was a problem hiding this comment.
successful status != successful operation
a successful status check means the metadata API returned a valid response, it can still be InProgress.
The errors caught in the catch block are exceptions thrown by jsforce (network errors, parsing failures, etc)
cristiand391
commented
Jan 6, 2026
| error: e, | ||
| count: this.errorRetryLimit, | ||
| }; | ||
| return { completed: true }; |
Member
Author
There was a problem hiding this comment.
completed: true to signal the polling client to stop
cristiand391
changed the title
cristiand391
commented
Jan 6, 2026
Increase the default consecutive error retry limit from 25 to 1000 to provide more tolerance for intermittent network issues during long-running deploy/retrieve operations. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
[skip ci]
Contributor
|
Looks good to me! Tested this by disconnecting wifi mid deploy/retrieve.
|
iowillhoit
approved these changes
Jan 7, 2026
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What does this PR do?
Updates
MetadataTransferclass to track errors during retrieve/deploy polling.The polling is done using sfdx-core's
PollingClientwithout specifyingretryLimit, so we have it run until a timeout happens:https://github.com/forcedotcom/sfdx-core/blob/5564069767b85a96e73f8bf88dbdd3d7e4b5da03/src/status/pollingClient.ts#L131
During a retrieve/deploy with
sf, if the metadata api starts to return one of the retryable errors constantly (backend or metadata issue?) the polling keeps going until a timeout (--waitflag value) and then throws a generic "client timed out" error without much info about the real issue.This PR adds an error tracker to count consequent errors during polling checks, allowing to:
What issues does this PR fix or reference?
@W-18203875@
Functionality Before
SDR not defining a retry limit would cause consequent api errors to be retried until timeout and throw a generic timeout error
Functionality After
SDR tracks consequent errors, throws after 25 consequent errors during pollling and allows to customize the limit via env var.