Analysis methods
Besides comparing a subset of spans with all the rest in the global timeframe, it gives you ways to:
- Analyze attribute values: analyze how attributes and their values cluster across all the available spans. Maybe you are not looking to troubleshoot issues, but rather understand the reality of your data.
- Compare with the other spans in the same timeframe: this method is very good to analyze outliers, especially in terms of duration.
- Compare with everything before the selection in the global timeframe: this is an excellent way to analyze the first occurrence of spikes or error clusters.
Triage is in Beta
This is the first iteration of Dash0 triage. It is available in open beta for Tracing, which is the signal with the largest amount of metadata on average.
At time of writing this blog, we know of various cosmetic bugs and UX details that need improvement, but we don’t want to wait to learn with and from you and your use cases. Also, we are working hard on the performance (sifting on a lot of data the way Triage does requires optimizations), and until we are happy with the outcome, you can use Triage only on timeframes of up to 24 hours.
Triage will also come for Logging and Metrics when we feel we have confirmation through your feedback that we nailed user experience. Indeed, expect the user experience to be tuned and improved in the next few days and weeks.