What's new — October 2026

We've shipped 42 improvements across InfraNest — from automating your infrastructure tasks to making your status pages work harder for your team.

✨ New

• New domains can now arrive with DNS records pre-filled from templates — choose a mail provider, security baseline or static site starter set while buying, see exactly what will be created, and everything lands in your DNS provider's zone as soon as the registrar confirms.
• Alerts, domain notices and status-page emails can now carry your own logo, accent colour and sender name instead of ours, set in Settings, Organization, Brand, with a preview showing how customers will receive them.
• When registering a domain you can now pick one of your connected DNS providers — Cloudflare, Route 53, Hetzner, DigitalOcean — and we create the zone in your own account there and point the domain at it as soon as registration is confirmed.
• Automations can now file Linear issues — connect once and any rule can open a ticket on the team you choose with details already filled in, selecting from your real Linear data instead of hunting for IDs.
• The Security Advisor now reads five Cloudflare settings beyond DNS records — SSL mode, minimum TLS version, HTTPS redirect, HSTS and development mode — flagging risks like Flexible SSL that only half-encrypts your connection.
• Drop Catch now recognises that .ai, .ca and .cz release names to auction first, not straight to registration, factors that into the odds, and times its final checks for when registration can actually succeed.
• Domain watches now tell you all three ways they can end — the owner renewed after all, the name became registrable at standard price, or the registry data went stale — and automations can act on renewals with the new expiry date included.
• Automations can now resize a volume or rename one — a rule can grow a filling disk or keep volume names in step with changes elsewhere.
• Servers can now be renamed — click the name at the top of the page and the change is made at your provider first, so it shows in their console too and the audit log records what it was before.
• Alert destinations now report their own health — after repeated failures they are marked Not delivering on your Alerts page and owners are told by email and in the bell whether the service refused credentials or is just having a bad hour.
• Your identity provider's groups can now decide roles here — map each group to a role and people land on the right one when they sign in, with changes following your provider unless you override by hand.
• Your team can now sign in with Okta, Microsoft Entra ID, Google Workspace or any OpenID Connect provider — verify your email domain and people arriving from your provider are recognised straight away or sent for approval.
• Automations can now file Notion tasks — connect once and any rule can add a task to your database with details filled in, picking your own columns by name and type instead of a fixed set of fields.
• Automations can now open Jira issues — connect once and any rule can file a ticket in your chosen project with details filled in, and comment, move or link the issue.
• Status pages now have a Team view showing services grouped as customers see them, with last 30 days, current state and check freshness, plus an on-call contact and runbook link on each service for whoever picks up at 3am.
• Status pages now have a Layout setting for one column or two — two columns puts services on the left with maintenance and subscribe form beside them, folding back to one column on phones.
• Every incident on your status page now has its own link — paste it into a support ticket and the customer lands on that incident with what it affected, when it started and every update, kept after the incident resolves.
• Status pages now have an Appearance section — pick a colour scheme including Auto that follows device settings, spacing, history depth, a header banner, and switches for every section you do not need.
• DNS records pointing at a deleted server are now flagged as errors, firewall rules allowing old addresses are flagged, and monitors watching them say so — each names the server and when it left, and the delete box shows what in InfraNest points at that address.
• A server reading that is genuinely climbing now opens the page with a line saying which one, how fast and roughly when it matters — 'Up 18d 4h, but CPU reaches its 90% alert line in about 6 days' with a button to the chart.
• Before a domain drops we now check every registrar you armed a watch with against the extension's price and warn you while there is still time to top up.
• Two new alerts under Monitoring → Alert defaults for servers with the agent — a reboot is pending tells you once when a restart is needed, and security updates are waiting fires when updates go uninstalled for a day. (details)
• Power Automate joins the automation integrations — start a flow when a domain expires, a monitor goes down or a certificate lapses, and every outgoing endpoint now shows the schema of what we send.

⚡ Improved

• For .uk domains Drop Catch now uses Nominet's published drop times instead of estimates, and gradually backs off checking stale names instead of hammering them forever, making 'dropping soon' alerts actually mean soon.
• The help centre now opens on ten categories with descriptions instead of 168 articles in a scroll, and every page header carries a help button that opens the article for that page.
• Delete protection on volumes and networks is now mirrored to your provider, so the lock holds even in your provider's own console.
• Sign-up now refuses email addresses whose domain cannot receive mail at all, and suggests corrections for likely typos at common providers.
• Security keys now count as two-factor authentication — if your organisation requires it, two passkeys or one alongside an authenticator app satisfy the requirement.
• The domain basket now checks whether a name can be bought before the button is live — it catches missing contact info, extension requirements, registrar support and balance problems up front instead of by email after the order fails.
• When a registrar or cloud provider refuses an order, you now get a sentence you can act on and a button to the exact page that fixes it, with their raw error still quoted underneath.
• You can now search for, buy and watch domains with accents or in other scripts — münchen.de, café.fr and пример.рф work typed exactly as you would write them, and a search that comes back empty says which part could not be used.
• Connecting an identity provider now walks you through setup step by step — showing what to do at your provider, checking your email domain, signing you in as a test before anything switches on.
• Each PagerDuty account now has a Send test incident button proving both directions at once, and Check setup compares what InfraNest created with what is there now before making any changes.
• Silencing a server now lets you pick how long — an hour, four hours or a day — and it always ends by itself, with rules still being checked so anything still wrong is reported when it ends.

🐛 Fixed

• Drop Catch now reads actual deletion dates from the registry instead of guessing, fixing timelines that were sometimes off by nearly a month, and stops treating pending-delete statuses the same way across extensions that handle them very differently.
• Proxied Cloudflare records and DigitalOcean's 30-second TTL can now be saved again — we stopped refusing the TTL values that the providers themselves offer.
• A registrar hiccup no longer leaves a new domain pointing nowhere — brief failures when pointing a domain at its DNS zone are now retried instead of being written off forever.
• Records can now be edited on a zone whose root uses a flattened CNAME, since ALIAS records are designed to sit alongside other records.
• The MX 0 . record declaring that a domain receives no email can now be saved, letting you turn away mail immediately instead of falling back to your web address.
• The Unsubscribe button that Gmail and Yahoo draw from our bulk emails now actually works instead of being rejected.
• A refused domain order now explains itself by email in your own language with a button to fix it, and the alert for a domain that left your account is no longer an empty copy-paste error.
• The monthly uptime report now covers the whole month instead of dropping the most recent weeks — in-progress months show what has been measured so far, finished months count every day, and the number stops moving when the month ends.
• Dutch and German now address you informally throughout the app, and French formally, matching how the emails we send address you.