Propagate additional trace data into AWS requests on Lambda #549
Conversation
.gitignore
Outdated
| @@ -5,3 +5,4 @@ docs | |||
| .nyc_output/ | |||
| *.lcov | |||
| dist | |||
| *package-lock.json | |||
There was a problem hiding this comment.
I believe we need the package-lock.json file to keep track of the versions of the packages we use, and it is also one of the files that is modified when we are preparing for a new release. Is there a reason for adding it to .gitignore?
There was a problem hiding this comment.
I see. I thought it was auto-generated and need not be checked in. I will remove it from .gitignore.
packages/core/lib/utils.js
Outdated
| @@ -162,33 +162,37 @@ var utils = { | |||
| */ | |||
| populateTraceData: function(segment, xAmznTraceId) { | |||
| logger.getLogger().debug('Lambda trace data found: ' + xAmznTraceId); | |||
| var data = utils.processTraceData(xAmznTraceId); | |||
| var traceData = utils.processTraceData(xAmznTraceId); | |||
There was a problem hiding this comment.
nit: use let instead of var
packages/core/lib/utils.js
Outdated
| var key = pair[0].trim(); | ||
| var value = pair[1].trim(); | ||
| var lowerCaseKey = key.toLowerCase(); | ||
| var reserved = reservedKeywords.indexOf(lowerCaseKey) !== -1; |
There was a problem hiding this comment.
Is it necessary to create a new lowerCaseKey variable instead of converting the key & value to lowercase using .toLowerCase?
nit: use let instead of var.
There was a problem hiding this comment.
I don't want to mutate the casing of additional trace data since it may contain a case-sensitive key. So I created a separate variable lowerCaseKey from key as the former is used in multiple places below for existing logic.
Also, I don't think we need to change the casing of the value anywhere.
This PR allows the SDK to parse any additional key-value pair supplied in the X-Ray trace header. And then these additional key-value pairs are injected into instrumented AWS requests made from a Lambda function.
The purpose of this PR is similar to what I did for the Python SDK. The difference there was that the segment already had functionality to save the original trace header. Here, I added a new property on the segment
additionalTraceDatawhich stores only the additional key-value pairs instead of the complete trace header.By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.