I think this might be related to the post from yesterday about Vertex AI capacity in the eu multi-region, but we’re consuming via Gemini Enterprise (formally Agent Space).
It’s gradually got worse over this morning, but as of about 10am, we’d get nothing but timeouts or users just sat with “Working on it…” for ~20 mins, specifically for any chat with a context over about 250K tokens.
<250K tokens and it’s slow but usable.
I’ve got a P1 open with Google Support to see if they’ll officially acknowledge an issue, but I wanted to know. Anyone else seeing this?