Few years back, I used to monitor performance using Opbeat APM. It was quite easy to set up and track performance bottlenecks. It was later acquired by elastic and made the entire ELK stack mandatory to use APM.
Since then, I have been searching for an APM tool that just works without too much hassle. I have tried out APMs like Datadog, Elastic APM, Django-silk, and a few others.
I prefer agentless performance monitoring tools instead of APM tools which require additional agents to be run on servers.
I have been using Sentry for a long time and recently came to know about Performance. @BYK has helped me in resolving issues related to Performance.
Enabling performance was literally a single line change. Within a couple of days, I have found most of the performance issues with just a couple of filters in Discover. I think Sentry Performance on cloud is really worth paying for.
Sentry is a market leader in error tracking and I believe it will soon become the market leader in Python/JS APM space.
Here are some suggestions related to Performance and Sentry in general.
Overhead: It is better if some documentation is added regarding the overhead. See this for more details. https://github.com/getsentry/sentry-python/issues/668
Impact factor: I couldn’t figure out how to do simple arithmetic operations in discover. Not sure if it is supported yet. How to show calculated columns in Sentry performance?
Discover: In a recent webinar about metric alerts, Sentry team mentioned Discover as a separate pillar in Sentry. However, there were no webinars about Discover on the youtube channel. It would be great if you can plan a webinar on Discover. https://www.youtube.com/channel/UCP_sweTJ0DQmMlpvNSOpEJg
Tags: I wanted to go through all the blog posts related to performance on the Sentry blog. However, there were no tags for blog posts and I ended up searching through archives to find relevant posts. It would be great if you can add tags to the blog.