| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
| |
Lots of OpenStack projects are moving from direct usage of engine
to enginefacade module usage with its _TransactionContextManager.
Sessions for the context are created via .using() method of this
manager.
Change-Id: Ie517524a10713e09bb2491f2d76b651b20ecd4b6
|
| |
|
|
| |
Change-Id: I83aa97b3caadb053fddc76088d19ed5110035c67
|
| |
|
|
| |
Change-Id: Id335bfdf49ed88edd14e896a48c22b11113600c8
|
| |
|
|
|
|
|
|
|
|
| |
Use before_cursor_execute and after_cursor_execute instead of
before_execute and after_execute.
1) Migration with custom types will work now (even if tracing is enabled)
2) We don't need to render 2 times SQL expressesion => less overhead
Change-Id: I7c6b6909ce0f15a69ce5caad544e1351a647b472
|
| |
|
|
|
|
| |
It's easy for end users to find all public methdos when they are on top.
Change-Id: I0b80ea745d42dfec8523730a1f0410e2e8a1d388
|
| |
|
|
|
|
| |
Add "_" to private things, so user won't access them directly
Change-Id: Iab7d42c3b0a3f5ac3952b2be5c688ac939761701
|
| |
|
|
| |
Change-Id: Ieb9181a2e997e5e3ea02fd91d2b85aa54a4b83be
|
| |
|
|
|
|
| |
Engine that is used in migraiton shouldn't be traced.
So add built-in support for disable tracing.
This simplifies a lot code in OpenStack projects
|
| |
|
|
| |
Mostly this patch just covers by UTs sqlalcehmy module
|
| | |
|
| | |
|
|
|
To use it in your code you should add 2 sqlalchemy engine event listeners
import sqlalchemy
from sqlalchemy import event as sa_event
from osprofiler import sqlalchemy as sa_profiler
engine = sqlaclhemy.create_engine('your_db_host')
sa_profiler.add_tracing(sqlalchemy, engine, 'your_service')
|
