How do you use Sentry?
Sentry Saas (sentry.io)
Version
2.59.0
Steps to Reproduce
Python: 3.12
Django: 5.1
We upgraded the Python SDK from sentry-sdk[celery,django]==2.20.0 to sentry-sdk[celery,django]==2.59.0 and observed a large increase in memory usage, from 50% to 85%.
The screenshot below shows memory dropping back down after we reverted the upgrade, on a Django app with no traffic.
In both versions, only error tracking is enabled (no performance/tracing, no profiling).
Expected Result
Memory usage remains stable after upgrading sentry-sdk from 2.20.0 to 2.59.0, with no significant change in baseline RSS for a Django app under the same load (or no load).
Actual Result
After upgrading sentry-sdk[celery,django] from 2.20.0 to 2.59.0, baseline memory usage on our Django app increased from ~50% to ~85%, even with no incoming traffic. Memory returned to previous levels immediately after reverting the upgrade.
How do you use Sentry?
Sentry Saas (sentry.io)
Version
2.59.0
Steps to Reproduce
Python: 3.12
Django: 5.1
We upgraded the Python SDK from
sentry-sdk[celery,django]==2.20.0tosentry-sdk[celery,django]==2.59.0and observed a large increase in memory usage, from 50% to 85%.The screenshot below shows memory dropping back down after we reverted the upgrade, on a Django app with no traffic.
In both versions, only error tracking is enabled (no performance/tracing, no profiling).
Expected Result
Memory usage remains stable after upgrading sentry-sdk from 2.20.0 to 2.59.0, with no significant change in baseline RSS for a Django app under the same load (or no load).
Actual Result
After upgrading sentry-sdk[celery,django] from 2.20.0 to 2.59.0, baseline memory usage on our Django app increased from ~50% to ~85%, even with no incoming traffic. Memory returned to previous levels immediately after reverting the upgrade.