When will Llama Parse be back up? I

just updated the status page

currently investigating

looks like it should be back up

albeit with some delay as the queue burns down

hmm possibly optimistic on that fix, appears to still be degraded

Status page is updated though

https://llamaindex.statuspage.io/

Ok, should be back now

@Logan M it seems to be down still

@galvangjx seems to be working for me at first glance -- any specific settings you tried?

multimodal parsing with instructions

the UI is working for me but not through python

also reading document stored in azure blob storage

the jobs are taking too long to complete (more than 10 mins)

@galvangjx it seems to work fine for me? (both multi-modal and without, and it runs decently fast)

Are you using your own api key for the multimodal model? Maybe its rate limiting?

im using my own openai api key with available credits. im just starting out my day so the rates should not be limited. the parsing status is stuck at Started parsing the file under job_id 1012293f-44f7-4d9f-ba4b-6959175b3bf9

Looking at logs, seems like
a) gpt-4o may be slow right now
b) your parsing instructions seem to be huge. Maybe try removing or toning that down and see if it helps?

But something may be bugged too, it's been waiting for 11 mins without a timeout. That can get addressed in the next release if we can track that part down

But I think optimizing your instructions will help a bit

are there any guidelines for writing the parsing instructions?

for the size of the parsing instructions - i have been sending them and its always working fine till today