This commit is contained in:
@@ -37,6 +37,19 @@ DB INDEXES / PERFORMANCE
|
|||||||
- Useful for learning: https://www.datadoghq.com/blog/postgresql-monitoring/
|
- Useful for learning: https://www.datadoghq.com/blog/postgresql-monitoring/
|
||||||
-
|
-
|
||||||
|
|
||||||
|
DB CONNECTIONS:
|
||||||
|
https://news.ycombinator.com/item?id=18425923
|
||||||
|
wild_preference 2 hours ago [-]
|
||||||
|
The other half of this is correctly using the connection pool in the application layer to begin with.
|
||||||
|
One of the most common performance errors I see in random http applications is grabbing db connections for too long, like in middleware, or too aggressively in parallel queries.
|
||||||
|
reply
|
||||||
|
sb8244 2 hours ago [-]
|
||||||
|
This is so true and easy to overlook until a post mortem. A great exercise is to audit how connections are maintained in your application, before it's too late.
|
||||||
|
reply
|
||||||
|
macintux 1 hour ago [-]
|
||||||
|
I'm reminded of a systems failure described in, IIRC, the O'Reilly book "Release It!". Java ODBC exception swallowed resulted in eventually running out of connections.
|
||||||
|
Very tough problem to troubleshoot without extensive monitoring.
|
||||||
|
|
||||||
|
|
||||||
DOCUMENTATION TODO's
|
DOCUMENTATION TODO's
|
||||||
|
|
||||||
|
|||||||
Reference in New Issue
Block a user