idea0rbit

5 ms·

1. 1.▲
   by idea0rbit 15d ago

   I’m building TraceDB[1] because I think observability needs an overhaul. We pay a premium to index millions of near-identical spans, sample away the rare ones that might explain an outage, and still have to piece together what failed, where
2. 2.▲
   by idea0rbit 7mo ago

   I’m building an observability system that tries to surface answers instead of making people dig through huge amounts of raw telemetry. The basic idea is that when one failure fans out across 20 services, you often end up with 20 alerts and
3. 3.▲
   by idea0rbit 1y ago

   You are absolutely right. What could an alternate design and business model look like?
4. 4.▲
   by idea0rbit 1y ago

   both. In general I think most observability systems are broken. This article captures the sentiment pretty well https://www.linkedin.com/pulse/observability-broken-its-time...
5. 5.▲
   Ask HN: Why is observability so broken, and what can fix it?

   6 points by idea0rbit 1y ago|5 comments
6. 6.▲
   by idea0rbit 1y ago

   After a decade building large-scale systems at Google, Datadog, and Meta, I’ve noticed the same pattern repeat: observability keeps getting louder, costlier, and less useful. We’re drowning in telemetry but starving for insight. The industr
7. 7.▲
   Why is observability so broken?(omji.ai)

   2 points by idea0rbit 1y ago|1 comments