search

freelawproject / courtlistener

A fully-searchable and accessible archive of court data including growing repositories of opinions, oral arguments, judges, judicial financial records, and federal filings.
https://www.courtlistener.com
Other
550 stars 151 forks source link

Elastic Optimizations: Round Three #4631

Open mlissner opened 3 weeks ago

mlissner commented 3 weeks ago

We finally have a baseline of what our performance looks like on average. It's not horrible, but I think there's room for improvement. The hard part is figuring out what's wrong, but now we can measure each step we take and see what it does.

If this is the correct definition of types: 

o | Case law opinions
r | List of Federal cases (dockets) with up to three nested documents. 
rd | Federal filing documents from PACER
d | Federal cases (dockets) from PACER
p | Judges
oa | Oral argument audio files

here are some results:

Image

A few ideas:

What else?

I think this can wait until our new dev ops person comes on and it should be a good one for them. Eventually I'd also like to get this info into Grafana somehow, so we can see a chart of these speeds, but, well, all in due time!

Originally posted by @legaltextai in #4209

mlissner commented 3 weeks ago

I'm adding more nodes over in https://github.com/freelawproject/infrastructure/issues/193. When that's complete and a week has gone by, we should re-run our averages.