Closed mjeammet closed 4 weeks ago
@mjeammet I've finally understood what the problem is (thanks to @NathanVss 's help).
We use the DockerflowMiddleware
whose job is precisely to add a log for each request. So we there is an exception, we catch it and the middleware also adds a log : twice in Sentry.
I think the proper solution is to ignore the request.summary
logger with:
from sentry_sdk.integrations.logging import ignore_logger
#...
sentry_sdk.init(
# ...
)
# Ignore the logs added by the DockerflowMiddleware
ignore_logger("request.summary")
With that, we won't have duplicated logs anymore and we will be able to keep the default integrations :)
Purpose
a previous commit turned sentry's default_integrations to false to remove duplicates on monitored issues. Thing is, duplicates were appearing because we actually logged + raised the issue where only one of these methods is necessary. Better restore Sentry's default parameters and try to clean our code of duplicate logging
Proposal