Critical Feedback: Gemini 3.1 Instability, Agent Manager UX Downgrade, and Vague Changelogs

Google Antigravity
1

We’ve been integrating Antigravity into our workflows for several months with great results, but recent updates are seriously compromising our operations. With the rollout of Gemini 3.1, the experience has shifted from seamless to a constant loop of error management.

I want to raise three critical points that require the development team’s attention:

1. Instability and Critical Errors (Gemini 3.1)

Since Gemini 3.1, the failure rate has spiked. We are constantly hitting the “Agent terminated due to error” screen, forcing manual retries or new conversations. This is highly disruptive when running multiple agents in parallel, as half of them end up crashing without a clear cause.

2. Management UI Degradation (Agent Manager)

Removing the “Inbox” and lacking a chat search feature in the Agent Manager is a massive step backward for usability. Currently:

  • Navigating through tabs to monitor agent status is incredibly tedious.

  • Losing centralized visibility (the Inbox) makes spotting the 3.1 errors mentioned above a slow, manual chore.

  • Request: Please reinstate the Inbox and add search functionality to help us manage high volumes of agents efficiently.

3. Vague Changelogs & Silent Quota Reductions

The current changelog is far too vague. For those of us relying on this tool professionally, generic notes aren’t enough. We need detailed logs specifying:

  • Fixes: Exactly which bugs were squashed.

  • Deprecations: Why features like the Inbox are removed.

  • Infrastructure changes: Unannounced reductions in usage quotas directly hurt our technical planning.

Antigravity is a fantastic tool, but the current instability of the 3.1 model, combined with UI downgrades and poor communication regarding updates, is severely impacting our productivity.

Is anyone else experiencing this high frequency of agent crashes lately?

5 Likes
2

Yeah… there is another post with many people complaining about the agents error situation: https://discuss.ai.google.dev/t/our-servers-are-experiencing-high-traffic-right-now-please-try-again-in-a-minute
I don’t know what is happening with Antigravity lately :frowning:

1 Like
3

To Google as a whole. AI Studio has been down for more than 8 days already. Gemini chats were down, etc.

They’re doing some crappy updates. It looks like no QA has been done prior. They’re just rolling out a raw product full of bugs and errors.

Vibe coding is dead today. AI studio down, Antigravity down, Claude down.

I’m amazed how they fail to correct the issues for many days. They’re like playing in a kindergarten, pushing out broken/untested updates and killing the workflows of thousands of people.

1 Like
4

agent terminated due to error is gemini 3.1 hitting the guard rails to quickly. talk to it and discuss what happened. you can keep talking in the same chat, you don’t have to start a new one.

1 Like
5

I doubt that… when I get that error, if I click on “Copy debug info” I get something like this:

Trajectory ID: x
Error: HTTP 503 Service Unavailable
Sherlog:
TraceID: x
Headers: x

{
“error”: {
“code”: 503,
“details”: [
{
“@type”: “type.googleapis.com/google.rpc.ErrorInfo”,
“domain”: “cloudcode-pa.googleapis.com”,
“metadata”: {
“model”: “gemini-3.1-pro-high”
},
“reason”: “MODEL_CAPACITY_EXHAUSTED”
}
],
“message”: “No capacity available for model gemini-3.1-pro-high on the server”,
“status”: “UNAVAILABLE”
}
}

503 Service Unavailable. This has been happening too much recently.

1 Like
6

sounds like you’re out of quota

7

Thank you for bringing these concerns to our attention. Please be assured that I have shared your feedback with our internal team for further review.
We appreciate your continued patience as we work to enhance the Antigravity experience.

1 Like
8

Nah… if I click retry it works, it has nothing to do with the quota.

1 Like
9

there are exactly 2 things that a MODEL_CAPACITY_EXHAUSTED error can be - you’re out of quota/credit or something is overloaded. this isn’t an API call for you, correct?

10

But this is what is happening to everyone today. It’s the post I linked in my first message: https://discuss.ai.google.dev/t/our-servers-are-experiencing-high-traffic-right-now-please-try-again-in-a-minute
Hopefully it gets sorted soon.

1 Like
11

Anti Gravity has come a long way. Previously I ran into latency issues and slower instruction processing, but this latest experience was smooth and fast. Response time is dramatically improved, and the orchestration between agents feels much tighter. Big step up in performance.

1 Like