Changelog
Page 3 of 3.
Referrals now qualify on real connected uptime, for both operators
Referral qualification now counts real connected uptime — 1,440 hours on a single node — and both the referrer and the operator they referred have to reach it. Everyone's hours start again from zero, so the earliest any referral can qualify is around mid-November 2026.
Passkey sign-in to the portal works again
Signing in to the portal with a passkey failed for several weeks, because a security certificate one of our sign-in domains depends on had expired. Password sign-in kept working throughout, as did nodes and the sites they serve.
The dashboard's Docker setup guide link now opens the guide
On the dashboard's node install card, the Docker tab's "Docker setup guide" link led to a page that didn't exist. It now opens the Docker section of the installation guide.
Deleted workload secrets stop working across the fleet within five minutes
The control plane kept every workload secret it had loaded in memory until it restarted, so deleting a secret didn't stop running edge nodes decrypting that workload. Loaded secrets now expire after five minutes.
Deploy tokens keep every role you select
A deploy token created with both roles ticked lost its workload role, so every deploy with it was refused with "Insufficient permissions". New tokens keep every role you select. If you have a token affected by this, create a new one.
Metro leaderboards show a podium, each node's score breakdown, and work on phones
The metro leaderboards now put the top three nodes on a podium, let you expand any node to see how its Performance Index is made up, and turn into cards on a phone. Two score components that were labelled wrongly now carry the right names.
New operators appear on the Community page within minutes
A new operator's first node went live immediately, but the Community page could take up to an hour to list it, and until then said they had no nodes. It now catches up within a few minutes, and says your node is syncing while it waits.
The control plane records node state far less often
The control plane now saves each node's peer-discovery details only when they change, instead of on every heartbeat, cutting its routine database writes about sevenfold. Delivery is unchanged; two rare recovery cases behave slightly differently.
Card titles on the Workloads page are readable again
A styling conflict made some portal headings near-black on a dark background, most visibly the Content Delivery card title on the Workloads page. All 18 affected elements are fixed, and a test stops the conflict coming back.
Sites deployed from CI now appear in the portal
Sites published from CI or the command line served normally but left the portal's Files and Deployments tabs empty, and their traffic was missing from project analytics. New deploys now show up in the portal just as portal deploys do.
Docs and Support are one click away in the portal
The portal's top bar has a new Resources menu with links to Docs and Support. It opens the guides for whichever view you're in, Node Operator or customer, in a new tab. Available on desktop.
A new account menu and an always-visible view switcher in the portal
Your profile, account settings and sign-out now sit in a menu under your name, and the switch between the Node Operator and customer views is always shown. The Refer page also links the referral program terms directly.
Dashboard cards now open Earnings, the Leaderboard and your profile
The Earnings & Payouts, Local Leaderboard and Performance Index cards on the dashboard were dimmed and marked "Coming soon", though all three features were already live. Each card is now a link to its page.
Refer other operators from the portal
Operators now have a personal referral link and a Refer page in the portal. When someone you refer keeps a verified node online for 60 days, you both earn a bonus. There's no cap on referrals.
Sites stay up when an edge node's credentials are rejected
When the control plane rejected an edge node's credentials, the node kept retrying them, and sites it served returned "domain not configured". Nodes now re-authenticate within seconds, and the visitor who triggers recovery still gets their page.
Static sites are now cached by browsers and crawlers
Sites served from their own domain were sent with no-cache on every file, so each visit re-downloaded everything. Build assets are now cached for a year, and pages revalidate in a single cheap round trip.
Healthy nodes no longer show single-digit uptime
When a node's activation date was missing, its uptime was measured against a full 30 days, so a node with two perfect days showed 6.67%. Uptime is now measured against the time we actually observed the node.
Installing the edge node on Windows now works end to end
The Windows installer failed to download the edge node, and beyond that registered a service that couldn't start while still reporting success. Installing, starting and uninstalling now all work, and failures are reported as failures.
Subscribe: all products · CDN · Edge Node · CLI · Portal