Every custom domain now has a redirect target dropdown and a 301/302 status code selector. Point an old or alternate domain at your primary one and the platform handles the redirect for you, with a built-in bypass so SSL certificate renewals are never blocked by the redirect.
Before: Redirecting one domain to another meant handling it in your own application code or a manual ingress tweak. After: Pick the destination domain and status code from the settings tab, and the redirect is applied automatically.
The monitoring tab has been rebuilt around always-on metrics with 24 hours of retained history, so you no longer have to enable monitoring or extend a time window to see data. Large CPU and memory charts plot every service together, with a 1h/6h/24h range switcher, per-service filters, and a tooltip that lists each service's value at any point in time.
Before: Metrics had to be manually enabled, retained only 30 minutes of history, and showed a single combined chart. After: Metrics are always recorded for the last 24 hours, with multi-service charts you can filter and zoom by time range.
A new managed services secrets modal on the application settings tab gives you copy-ready environment variable blocks for your team's managed databases and object storages. Toggle between Laravel and generic naming, and reveal credentials on demand so secrets only load when you explicitly ask for them. Object storage access keys also gain an env snippet action on the Users tab.
Before: Wiring a managed database or object storage into your app meant assembling connection environment variables by hand. After: Open the modal, copy the ready-made block, and reveal credentials only when you need them.
Valkey services now have a toggle to emit their connection environment variables using REDIS_HOST, REDIS_PORT, REDIS_PASSWORD, and REDIS_URL instead of the VALKEY_ prefix. Because Valkey speaks the Redis protocol, your Laravel app and other Redis clients can read these straight from the environment with no remapping.
Before: Valkey injected VALKEY_ variables, so Redis clients needed manual remapping. After: Flip the switch and Valkey provides standard REDIS_ variables your app can use directly.
Managed databases gain an explicit toggle to allow connections from anywhere. An empty IP filter does not open access on its own, so this switch sets the 0.0.0.0/0 rule for you, and a notice on the connection tab makes it clear when public access is enabled.
Before: There was no clear way to allow connections from any IP; an empty allowlist was ambiguous. After: Flip the "allow access from all IPs" switch and see a notice confirming public access is on.
Deployments now read the ext-* requirements directly from your composer.json and add the matching PHP extensions to the build, on top of the existing package-name detection. The check runs at the start of every build, so newly added extension requirements are picked up without a failed build first.
Before: A missing PHP extension would fail the build, and you had to add it manually before redeploying. After: Extensions listed in composer.json are detected and installed automatically on each build.
Clients adding a payment method through a public payment link can now fill in their billing details on that same page. Company name, address, country, and VAT number are collected up front, so invoices come out correct without the team owner having to enter the details afterwards.
Before: Public payment links only collected a card; billing details had to be filled in separately by the team owner. After: Clients enter their own billing details while adding their card, in one flow.
A new database pricing page shows exactly what managed databases cost before you create one. Toggle between 1, 2, and 3-node setups to compare plans, and find starting prices on the managed databases feature page and the main pricing page.
Before: Managed database pricing was only visible inside the panel while creating a database. After: Full pricing is published on the website with a node-count toggle and per-plan breakdowns.
The website has been redesigned with a cleaner, more technical monochrome look built around a single brand accent, replacing the previous multi-color style across the homepage, feature pages, docs, blog, and changelog.
The enable quick deploy card now stays hidden until your most recent deployment has finished successfully, so you don't wire up git webhooks to an app whose build is currently broken. Once quick deploy is set up, its management cards stay visible regardless of later build results.
The custom domains list on the application overview is now capped in height with smooth scrolling and fade indicators, so a long list of domains no longer pushes the rest of the page down or interferes with canvas zoom.
CPU and memory charts and their tooltips now use proper dark colors in dark mode, fixing near-white grid lines and a white tooltip box that made values hard to read.
Creating a Statamic application now enables npm build commands by default, pre-creates the license environment variable, and reflects the correct storage, PHP, and pricing defaults so new Statamic apps are configured correctly out of the box.
Adding a domain that another application still claimed in the cluster used to pass validation and then fail with an unfriendly error at deploy time. The platform now checks the cluster, the authoritative source for claimed hosts, and rejects the domain up front with a clear message.
Applications with two or more custom domains were not getting their wildcard ingress generated, so subdomains returned a 404 even though the wildcard certificate had been issued. The wildcard ingress is now emitted regardless of how many custom domains an application has.
Team members saw a 'the selected social account does not belong to you' error when saving repository settings connected through the team owner's account. Repository settings now resolve against both the team owner's and the current user's connected accounts, and the dropdown suggests the matching account automatically.
Deleting an application volume now performs a real cleanup in the cluster instead of waiting for the multi-week orphan sweep, so you can immediately reuse a volume name like 'volume-1' without it failing because the old volume was still attached.
The DNS pre-check during domain add and save produced false negatives for Cloudflare-proxied domains, where the probe could be challenged or blocked. Domains are now accepted without the live DNS probe by default while keeping format and preview-domain checks intact.
The username and password fields for application basic auth no longer trigger the browser to autofill your panel login credentials, preventing accidentally saving the wrong values.
Suspended applications stayed on the usage page with no indication they were no longer consuming resources, making the cost breakdown confusing. They now show a suspended badge and an estimated cost if reactivated, kept separate from your actual running totals.