Replies: 1 comment 4 replies
-
Hi @lvca, One thing I would consider is introducing a type id. The type id can then be correlated to a taxonomy of issues. The benefit of this is it would be easier for Arcade to handle tuning. +1 on the JSON format since that can be piped to monitoring infrastructure which can parse and generate events from it. |
Beta Was this translation helpful? Give feedback.
4 replies
# for free
to join this conversation on GitHub.
Already have an account?
# to comment
-
Some errors could be stored in some JSONL files under a new database
report
directory. This would allow auto-recovery procedures and the user could access them.The most common error is when an index cannot be rebuilt for some reason (out of memory, out of space, etc) and ArcadeDB automatically drops it to avoid using a broken index for queries. Once dropped there is no procedure to retry recreating the index again. Also, the user couldn't see the errors/warnings in the log, so such issues are not found until the application becomes slow (no index).
I was thinking of a file like
databases/<mydb>/health.jsonl
(jsonl format is one json per line). Example:By using JSONL, both user and ArcadeDB can read and take action on that.
Also, there could be stored performance issues, like the lack of indexes with queries
Beta Was this translation helpful? Give feedback.
All reactions