Evaluating Vector Databases as AppCat services#177
Conversation
mikeshootzz
left a comment
There was a problem hiding this comment.
Without digging to deep into this: have you checked if AppCat exposes the necessary parameters to configure this properly?
Yes, there's actually not much to configure for PostgreSQL: for StackGres based installs the user needs to add the pgvector extension and then enable it for the database via SQL. In CNPG it's already installed so the user just needs to enable it. We need to document that though as mentioned in the ADR. |
mikeshootzz
left a comment
There was a problem hiding this comment.
It's good as.
I would maybe give a quick rundown of what would need to be configured (e.g. claim to add extension). This would just emphasize the "yes, it's that easy" a bit more.
At your discretion : D
|
|
||
| == Context | ||
|
|
||
| More customers require vector databases for various AI related workloads. |
There was a problem hiding this comment.
Are there any specific features required or did customers only inquire about vector databases in general?
| To provide the best services possible some of the most popular open source solutions are evaluated in this ADR. | ||
|
|
||
|
|
||
| Solutions being evaluated: |
There was a problem hiding this comment.
How do these compare performance and resource consumption wise?
There was a problem hiding this comment.
I did not do any benchmarking. Would also have been a pain since every solution has their own API...
Chroma and qdrant were basically negligible, with less than 100mb memory and 1mCPU. Milvus with all its bundled services was by far the heaviest on resources even in idle. I did not take any note of the concrete numbers for that though.
|
|
||
| == Consequences | ||
| We do not adopt any new vector database services into the AppCat service catalog. | ||
| Instead we make users aware about the vector capabilities of the existing services. |
There was a problem hiding this comment.
How would customers make use of these capabilities? I'm assuming there will still be some initial implementation needed by us to allow the use of it.
zugao
left a comment
There was a problem hiding this comment.
Rebase on master. The number now should be 0048
28c30b2 to
c04e0c3
Compare
Summary
Additionally, marked the Wazuh ADR as rejected.
Checklist
change,decision,requirement/quality,requirement/functional,dependencyas they show up in the changelog