apache / amoro

Apache Amoro (incubating) is a Lakehouse management system built on open data lake formats.
https://amoro.apache.org/
Apache License 2.0
748 stars 261 forks source link

[Feature][AMS]: Add overview page in AMS dashboard #405

Open zhoujinsong opened 1 year ago

zhoujinsong commented 1 year ago

Description

We need an overview page in AMS dashboard to disply the summary information and metrics of Arctic.

Use case/motivation

Add an overview page in AMS dashboard to disply the summary information and metrics of Arctic.

Describe the solution

Related issues

No response

Are you willing to submit a PR?

Code of Conduct

majin1102 commented 1 year ago

visual design for overview page: image including:

XBaith commented 1 year ago

Hi @zhoujinsong @majin1102 , have you considered stats for Iceberg table health? For example, estimate the time cost of a simple query, if it is greater than a reasonable time, then the table is probably an unhealthy table. Then Arctic can directly tell the users which tables are not suitable for reading, instead of giving us feedback when the query is too slow.

zhoujinsong commented 1 year ago

@XBaith , I agree that it will be very useful if we can tell users if iceberg tables are healthy. As far as I can see, the unhealthy conditions of the table may be:

And yes, Arctic is trying to help tables avoid the problems above.

XBaith commented 1 year ago

@zhoujinsong Correct. But for the native iceberg format, "eq-delete rate" is also an important indicator. We can calculate eq-delete rate by total-eq-delete-records / total-records. On the iceberg JMM benchmark, query performance may be unacceptable when eq-delete is greater than 0.6 (estimated value)