Sevalla Changelog #0089 Faster Deploys, AI, & CDN Cache

Establishing connection…

Signal detected.

Link established.

Baud here with a 56k Sevalla changelog transmission.

Three updates shipped over the last two weeks. One gives teams direct control over CDN cache behavior. One removes waiting during static site deployments. One brings AI-powered troubleshooting directly into the dashboard.

The common theme: less waiting, less guessing, and fewer reasons to leave your development workflow.

Because production teams shouldn't have to open support tickets, stitch together AI tools, or wonder whether a deploy actually finished.

Quick links to the updates:

• Clear caches without waiting on support
• Faster deployments with better visibility
• AI support that can troubleshoot your account

Clear caches without waiting on support#

Purge CDN cache in one click for static sites and object storage#

Production teams move fast. Infrastructure often doesn't.

You deploy an update, replace an asset, and suddenly users are still seeing old content. Now you're waiting for cache expiration or asking support to intervene.

That's unnecessary friction.

Static sites and object storage now include a dedicated Settings → Cache section where you can purge CDN caches yourself.

Use it immediately after a deploy or upload when fresh assets need to be served globally right away.

Static sites

Purge edge caches after deployments so users see updated assets immediately.

Object storage

Purge bucket caches after replacing or uploading files so applications stop serving stale objects.

The change takes effect immediately, and you'll receive confirmation once the purge is complete.

Available to anyone with update access to the service.

Why it matters

Production teams shouldn't have to wait for infrastructure to catch up with their code.

You made the change. You should control when it's live.

Faster deployments with better visibility#

Static site deploys are now 32.84% faster#

Deployments shouldn't feel like sending code into a black box.

We've optimized static site deployments across the platform.

Across 264 sites, average deploy time dropped from 82.51 seconds to 56.6 seconds.

That's a 32.84% reduction without changing your workflow.

Push code like you normally do. The speed improvement is automatic.

We also made deployment logs more useful.

Instead of staring at a silent upload step, you'll now see:

• Total number of files being uploaded
• A live progress bar
• Current file count and upload size
• A final deployment summary
• Clear indicators when uploads fail

Example:

⬆️  Uploading 1842 files...
   [████████████░░░░░░░░]  60% · 1105/1842 files · 24.3 MB
✅ Uploaded 1842 files (41.2 MB) in 18.7s

Why it matters

Developers don't need more deployment tooling.

They need deployments that are faster and easier to trust.

Waiting less is good.

Knowing what's happening is better.

AI support that can troubleshoot your account#

AI support agent now available inside the dashboard#

Most AI support experiences stop at documentation.

That's useful until something breaks in production.

Sevalla's AI support agent can go further.

You can now chat directly with an AI assistant inside the dashboard. It can answer platform questions and, with your permission, inspect services running inside your account.

The assistant can help investigate:

• Deployment status
• Error logs
• Database backups
• Troubleshooting workflows

You can also attach screenshots to conversations.

Enable it under:

Help → AI support

You'll review data processing terms and explicitly choose what the assistant can access.

Permissions work like API keys.

You can:

• Use predefined roles
• Create custom permissions
• Restrict access to specific resources

Only company owners and admins can configure AI support.

A few additional details:

• Powered by Anthropic Claude
• Data is sent to Anthropic only to generate responses
• Data is not used to train their models
• Enabling AI support creates a managed API key
• Disabling AI support immediately revokes that key
• Human support can take over conversations when needed
• Without account access, the assistant still answers documentation questions

Why it matters

Teams increasingly build with AI.

They're now expecting to operate production systems with AI too.

The goal isn't replacing engineers.

It's removing the hours spent jumping between dashboards, logs, and documentation just to answer simple operational questions.

Transmission complete#

Three updates that remove friction from the moments developers encounter every day: deploying, troubleshooting, and getting changes into production faster.

More improvements on the way.

Awaiting next transmission.

Link closed.