I built a tool that automatically generates bugfix PRs from production logs. Am I solving the wrong problem?
Hey Indie Hackers,
Like a lot of you, I hate debugging production outages.
A few months ago, I was paged at 3 AM for a failing database query. By the time I:
Traced the alert in Datadog,
Located the offending lines in GitHub,
Figured out the root cause (an unhandled null state in an aggregation query),
Wrote the patch and pushed a PR...
...over 2 hours of sleep were gone.
I realised that alert systems are great at telling you when things break, but completely leave the why and how to fix it to manual labor.
So, I spent the last few weeks building DevOrbit (https://www.devorbit.live). The premise is dead simple:
You paste a Trace ID (from Datadog, GCP, New Relic, etc.).
It maps the error log to the exact line of code in your repo.
It automatically proposes a production-ready, syntax-highlighted GitHub PR with the bugfix.
I also built a two-way Slack integration where you can just run /devorbit investigate <trace_id> directly in your team alert channels, and it drops the diagnosis there.
Here is where I need your help:
Observability and error monitoring is a notoriously crowded space (Sentry, Datadog, LogRocket). But they are mostly passive charts.
If you are a developer or team lead:
Would you trust an AI assistant to write your production bugfixes, even if you have full human-in-the-loop review before merging?
Is logs-to-remediation a real daily pain point for you, or is "log digging" just accepted as part of the job?
Would love your honest, unfiltered feedback on the flow!
