[SLING-11748] Improve logging output of HTTP retries in testing clients

• Type: Improvement
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: Apache Sling Testing Clients 3.0.16
• Fix Version/s: Apache Sling Testing Clients 3.0.20
• Component/s: Apache Sling Testing Clients
• Assignee: Andrei Dulvac
• Reporter: Marc Pfaff
• Created: 11/Jan/23 08:41
• Updated: 15/Jan/24 08:37
• Resolved: 13/Oct/23 09:39

Description

The ServerErrorRetryStrategy used in the testing clients logs when a retry condition is met. 

I propose to improve this output with more request/response details and retry condition details, in order to help troubleshooting test failures. 

Current output (example from the unit test):

Request retry needed due to service unavailable response
Response headers contained:
Header Date:Wed, 11 Jan 2023 08:36:43 GMT
Header Server:TEST/1.1
Header Content-Length:8
Header Content-Type:text/plain; charset=ISO-8859-1
Header Connection:Keep-Alive
Response content: TEST_NOK

Proposed improvement (example from the unit test):

Request retry condition met: [count=1/4], [expected-codes=[200]], [retry-codes=[500, 503]]
Request: GET /test/internalerror/resource HTTP/1.1 [Host: 127.0.0.1:32953, Connection: Keep-Alive, User-Agent: Java, Accept-Encoding: gzip,deflate, Authorization: Basic dXNlcjpwYXNz]
Response: HTTP/1.1 500 Internal Server Error [Date: Wed, 11 Jan 2023 08:39:59 GMT, Server: TEST/1.1, Content-Length: 8, Content-Type: text/plain; charset=ISO-8859-1, Connection: Keep-Alive, ]

 

Issue Links

• is cloned by: SLING-12096 Testing clients - don't log request headers
• links to: GitHub Pull Request #43