Closed
Bug 1171527
Opened 10 years ago
Closed 10 years ago
Make upload_symbols.py retry on 500 errors from the API
Categories
(Toolkit :: Crash Reporting, defect)
Toolkit
Crash Reporting
Tracking
()
People
(Reporter: ted, Assigned: ted)
References
Details
Attachments
(1 file)
We got a 500 ISE from the symbol upload API today, but it was transient and retrying would have worked. We don't currently retry on 500.
| Assignee | ||
Comment 2•10 years ago
|
||
I have a patch, I tested locally with a python httpd that alternates returning 500 and 200:
0:01.36 Uploading symbol file "dist/firefox-41.0a1.en-US.linux-x86_64.crashreporter-symbols-full.zip" to "http://0.0.0.0:48030/"...
0:03.10 Error: got HTTP response 500: Internal Server Error
0:03.11 Response body:
0:03.11 ====================
0:03.11 <head>
0:03.11 <title>Error response</title>
0:03.11 </head>
0:03.11 <body>
0:03.11 <h1>Error response</h1>
0:03.11 <p>Error code 500.
0:03.11 <p>Message: Internal Server Error.
0:03.11 <p>Error code explanation: 500 = Server got itself in trouble.
0:03.11 </body>
0:03.11
0:03.11 ====================
0:03.11
0:03.11 Retrying...
0:15.85 Uploaded successfully!
Assignee: nobody → ted
| Assignee | ||
Comment 3•10 years ago
|
||
bug 1171527 - Make upload_symbols.py retry on 500 errors from the API. r?gps
Attachment #8615944 -
Flags: review?(gps)
Comment 4•10 years ago
|
||
Comment on attachment 8615944 [details]
MozReview Request: bug 1171527 - Make upload_symbols.py retry on 500 errors from the API. r?gps
https://reviewboard.mozilla.org/r/10383/#review9123
::: toolkit/crashreporter/tools/upload_symbols.py:72
(Diff revision 1)
> + # 500 is likely to be a transient failure.
> + # Break out for success or other error codes.
> + if r.status_code != 500:
> - break
> + break
> + print_error(r)
While you are here, could you make sure we have a log event for every attempted HTTP request? As it stands, we don't have a good way of measuring 500 frequency from clients. If there's one thing I've learned from services it's that server logs only tell part of the story.
Attachment #8615944 -
Flags: review?(gps) → review+
| Assignee | ||
Comment 5•10 years ago
|
||
(In reply to Gregory Szorc [:gps] from comment #4)
> While you are here, could you make sure we have a log event for every
> attempted HTTP request? As it stands, we don't have a good way of measuring
> 500 frequency from clients. If there's one thing I've learned from services
> it's that server logs only tell part of the story.
You mean log something in the loop here for every request it makes?
Comment 6•10 years ago
|
||
(In reply to Ted Mielczarek [:ted.mielczarek] from comment #5)
> You mean log something in the loop here for every request it makes?
Yes.
| Assignee | ||
Comment 7•10 years ago
|
||
Comment 8•10 years ago
|
||
Status: NEW → RESOLVED
Closed: 10 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla41
Updated•10 years ago
|
Flags: needinfo?(ryanvm)
Comment 9•10 years ago
|
||
status-firefox40:
--- → fixed
Comment 10•10 years ago
|
||
status-firefox39:
--- → fixed
Updated•10 years ago
|
Flags: needinfo?(ryanvm)
Comment 11•10 years ago
|
||
status-b2g-v2.2:
--- → fixed
Comment 12•10 years ago
|
||
status-b2g-v2.1:
--- → fixed
Comment 13•10 years ago
|
||
status-b2g-v2.0:
--- → fixed
Comment 14•10 years ago
|
||
status-firefox-esr38:
--- → fixed
Comment 15•10 years ago
|
||
status-b2g-v2.1S:
--- → fixed
Comment 16•10 years ago
|
||
status-b2g-v2.0M:
--- → fixed
Comment 17•10 years ago
|
||
status-firefox-esr31:
--- → fixed
You need to log in
before you can comment on or make changes to this bug.
Description
•