slow LLM config init | Dagger | Page 1

crude cypress May 9, 2025, 3:40 PM

#

parallel helped 👍

earnest solar May 9, 2025, 5:13 PM

#

It’s intentionally more expensive cause we need a not-easily-brute-forceable hash of the plaintext to use as the cache key, but it should be adding overhead of milliseconds.

#

Are those secrets from env vars?

crude cypress May 9, 2025, 6:09 PM

#

earnest solar Are those secrets from env vars?

env vars which contain op:// values:

                DOTENV: |-
                    ANTHROPIC_API_KEY=op://RelEng/ANTHROPIC/API_KEY
                    GEMINI_API_KEY=op://RelEng/GEMINI/API_KEY
                    OPENAI_API_KEY=op://RelEng/OPEN_AI/API_KEY

#

oh so maybe it's fetching the plaintext twice for each? once to compute the hash, and then again when we call .plaintext?

#slow LLM config init