Ongoing Work & Feedback

[aldy505]

Hello there! 👋

Self-hosted is a big part of Sentry and has provided many benefits to companies and communities, helping them track errors and performance. I believe you already know what Sentry can do. Through this issue, we're aiming to be open and transparent about: (1) our priorities; (2) our backlogs; (3) the features you can expect to see in self-hosted Sentry; and (4) how to submit feedback.

If you want to submit feedback (not just asking for help or explaining what went wrong with self-hosted -- but also things that made you happy while maintaining self-hosted Sentry), feel free to comment on this issue!

For known existing backlogs, you can visit this GitHub Project: Self-hosted Sentry

---

What we are prioritizing

These issues are being worked on by us.

• Cleaning nodestore_node table #1808 -- Current nodestore lives in Postgres and is causing lots of disk storage issues. We're trying to move to S3 or filesystem-based implementation. A draft for S3 with Garage is available here as a blueprint feat: Use S3 node store with garage #3498
• Support ARM64 #1585 -- Sentry is suffering from GH runner exhaustion. So until the Linux ARM64 runners are out of beta, we're trying to hold it off
• uWSGI reports full listen queue #1573
• DockerHub reduced the maximum pulls per hour, causing people can't pull all the images properly #3593 -- Low priority, but we're starting to getting there
• Support Podman (and Podman Compose) #369 -- PR is already made: feat(install): Adds support for podman(compose) #3673
• Enable Dynamic Sampling #1866 -- It is long overdue

What features you can expect to land on self-hosted

• Add Uptime Monitoring #3304
• Add New Trace Explorer + Span-based Metrics #3592
• Add Ourlogs #3560
• https://github.com/getsentry/seer (AI/ML/Autofix features -- still a long way)

[edgariscoding]

This is awesome!

Using Sentry, specifically self-hosted Sentry has been a life changer at work. We're a small team and using self-hosted allowed us to dip our feet into proper error logging. Now we've been able to fix hundreds of tiny issues in our code that would not have been easily found with our old error logging solution, so our application is performing better than ever. Thank you!

I'm extremely excited for Uptime Monitoring as well as Logging. I had been tasked with finding a centralized logging solution for our multiple applications and I think I may just wait for Sentry Logging instead.