8 | | It happened after 220 api calls within the span of 7minutes, which probably means x5-10 database calls. |
9 | | |
10 | | This forces me to turn of connection pooling. |
| 8 | Further investigation shows the reason was a Threaded Middleware orm interaction, like this simplified code. |
| 9 | |
| 10 | class TimeMonitorMiddleware(object): |
| 11 | batched_logs: List[RequestLog] = [] |
| 12 | |
| 13 | def __call__(self, request: WSGIRequest) -> HttpResponse: |
| 14 | # ... |
| 15 | |
| 16 | # Save batched logs |
| 17 | if len(self.batched_logs) >= 50: |
| 18 | Thread(target=RequestLog.objects.bulk_create, args=(batched_logs,)).start() |
| 19 | |
| 20 | return response |
| 21 | |
| 22 | |