Every release,
in the open.

Small polish on the API key creation flow. The modal that shows you your new key once it is generated is no longer cramped, the copy button now actually sits next to the key (with a clear "Copied" confirmation), the key auto-selects on click, and a small usage snippet with a curl example is now shown so the next step is obvious.

• FixThe "API Key Created" modal copy button now sits inline next to the key input with a proper "Copy" / "Copied" state, instead of a tiny floating clipboard icon that could land outside the input.
• ImprovedClicking the key field now auto-selects the whole key so Cmd+C / Ctrl+C works instantly without reaching for the button.
• ImprovedModal now includes a curl example showing how to send the key as a Bearer token, with a link to the API docs.

Two small but annoying things fixed. Blog posts no longer show a brief hydration flicker on the published date for readers outside UTC, and the US check region now always returns a proper, actionable error code instead of ever appearing to hang.

• FixBlog post and blog list pages now render the published date in UTC regardless of your timezone, so the date stays consistent and no longer triggers a client-side rendering warning in the console.
• ImprovedUS region checks are now more resilient: if DNS or a specific request phase stalls, the region returns a clear error (DNS_ERROR / TIMEOUT) within seconds instead of ever getting stuck. Your incident timelines from the US node are now strictly more accurate.

Extending yesterday’s response-time correctness fix. The /monitors list page and its uptime / performance sub-views were still showing each monitor’s last single check in the response-time column. They now show the same rolling 24-hour average as the detail page, so the numbers are consistent wherever you look.

• FixThe response-time column on the /monitors list, /monitors/uptime, and /monitors/performance pages now shows each monitor’s rolling 24-hour average instead of its last single check, matching the detail page.
• ImprovedOrg-wide averages, fastest/slowest picks, and the response-time histogram on /monitors/performance now reflect 24-hour aggregates per monitor rather than a single-moment snapshot.

Follow-up to yesterday’s response-time fix. The Response pill in the header row next to Uptime and SSL now reads the same rolling 24-hour average as the "Avg Response (24h)" stat card below it, so the two numbers agree and the whole header row is consistently showing aggregates rather than the pill breaking from the pattern.

• FixThe Response pill in the monitor header now matches the "Avg Response (24h)" stat card value, so a single recent check no longer makes the header pill disagree with the card.

The "Avg Response" number on each monitor detail page was silently showing the response time of the most recent check, not an average, so it would swing between refreshes. It is now a real rolling 24-hour average computed from every check in the window, and the header pill next to your uptime shows the current-health "last response" separately.

• FixThe "Avg Response" stat on a monitor's detail page is now a true rolling 24-hour average, labelled "Avg Response (24h)" so the time window is explicit.
• FixThe header health pill now shows the response time of the most recent check, separate from the 24-hour average, so you can see the current health signal without it moving the historical number.
• ImprovedManual check, queue re-check, and auto-test after creating a monitor all update the right field now: the "last response" pill moves on a single check, the 24-hour average only updates when fresh monitor data is fetched.

Maintenance times entered in the form are now interpreted in your organization timezone, not your browser timezone, so 7:49 AM means 7:49 AM to you regardless of where you are sitting. The create dialog shows the active timezone plainly, and cards render every timestamp with an explicit offset suffix.

• FixMaintenance start and end times are now interpreted in your organization timezone on submit, fixing a silent offset shift that affected any org not set to UTC.
• ImprovedThe "Schedule Maintenance" dialog now shows the active organization timezone clearly (e.g. "America/New_York (EDT, GMT-4)") with a link to Settings if you need to change it.
• ImprovedMaintenance cards now render Starts and Ends in the timezone the record was created in, with an explicit GMT offset, so timestamps stay stable even if the org timezone is changed later.

Broader follow-up to 2.23.1. Maintenance windows now suppress every notification type during the active window, not just UI Changed. SSL expiry, DNS change, content change, performance degraded, and slow response alerts all respect the window. History (DNS snapshots, screenshots, content hashes) is still captured so you can review exactly what happened during maintenance once it ends.

• FixEvery alert type (SSL, DNS, content, performance, slow response, visual) is now suppressed during active maintenance windows, not just UI Changed.
• FixData collection (DNS snapshots, content hashes, screenshots) continues during maintenance so your change history stays accurate, only the notifications are muted.

A fix and a polish. Scheduled maintenance windows now correctly suppress visual-change (UI Changed) alerts, not just uptime alerts, so a planned site update does not flood your channels. The Maintenance page card is also redesigned with clear Starts / Ends / Duration cells, timezone, live countdown, and clickable affected-monitor links.

• FixVisual change (UI Changed) alerts are now suppressed during active maintenance windows, matching the existing behavior for uptime alerts.
• ImprovedMaintenance page card redesigned: explicit Starts, Ends, and Duration cells with timezone, live countdown ("Starts in 4h 12m" / "Ends in 32m" / "Ended 3d ago"), and clickable affected-monitor links.
• ImprovedIn-progress maintenance windows now visually stand out with an amber border and pulse indicator so it is obvious which window is live.

The fourth guide in the error-diagnosis cluster lands in the Knowledge Hub. A plain-language explanation of what DNS errors actually are, what the Chrome error messages really mean under the hood, how a DNS problem can surface as "your connection is not private," and step-by-step fixes for visitors, site owners, and gaming consoles (PS5, Xbox).

• NewNew Knowledge Hub guide: "What Is a DNS Error? Meaning, Causes, and How to Fix It." Covers the six DNS response codes you actually see (NOERROR, NXDOMAIN, SERVFAIL, REFUSED, FORMERR, NOTIMP), the Chrome error surface (DNS_PROBE_FINISHED_NXDOMAIN, ERR_NAME_NOT_RESOLVED, and friends), how DNS problems can look like SSL warnings, and a six-step fix tree for visitors.
• NewSide-by-side PS5 and Xbox DNS-error fix instructions with exact menu paths and recommended DNS servers.
• ImprovedThe existing "DNS server not responding" guide and the new "What is a DNS error" pillar now cross-link: the pillar covers the broad category, the narrower guide drills into the resolver-unreachable case.

A focused pass on visual change detection after a customer report showed the same underlying change being split into two separate incidents with fragmented timing. The lifecycle is now end-to-end accurate: an incident stays open until the page really returns to baseline or you accept a new one, duration reflects the real change window, and the incident record itself carries the server correlation that the alerts already showed.

• FixVisual incidents no longer auto-close after 4 hours. A visual incident now stays open until the page actually returns to baseline, you accept a new baseline, or you resolve it manually.
• FixIf a visual change persists after an incident was resolved (against the same baseline, within 24 hours), the previous incident is reopened instead of a new one being created. "Detected at" and "Was changed for" stay anchored to the original detection time.
• FixVisual incident records now include linked server name, server health snapshot, and the diagnosis phrase, matching what the alert channels (email, WhatsApp, Slack, Discord) already show.
• FixRecovery emails now always include a "Was changed for" row, falling back to "Unknown" when the original detection time is not available, matching WhatsApp behavior.
• FixResolved visual incidents now persist their total duration to the database, so historical reporting and incident views show the correct "Was changed for" value.

Shipped two more Knowledge Hub guides (504 Gateway Timeout, SSL Errors) to round out the core error-diagnosis cluster, and refreshed the three highest-traffic comparison and explainer posts with first-party data sourced from our live monitoring.

• NewNew Knowledge Hub guide: "What Is 504 Gateway Timeout (Meaning, Causes, and Fix)." Covers the timeout trio of 504 vs 502 vs 408, platform-specific diagnosis on nginx, Cloudflare, AWS, and Vercel, and the Cloudflare 504 vs 524 distinction.
• NewNew Knowledge Hub guide: "What Is an SSL Error and How to Fix It." Decodes the 10 most common SSL error messages, platform-specific diagnosis across Chrome, Safari, Firefox, Edge, and major web servers, with openssl diagnostic patterns.
• ImprovedTop three comparison and explainer posts now carry first-party data from our own monitoring instead of generic claims.

A deep 502 Bad Gateway troubleshooting guide written from a DevOps engineer’s perspective lands in the Knowledge Hub, plus mobile polish on the blog and better tap targets on tools and compare pages. Homepage caching also got a cleanup behind the scenes so return visits feel snappier.

• NewNew Knowledge Hub guide: "What Is 502 Bad Gateway and How to Fix It." Covers root causes, diagnosis on nginx, Cloudflare, AWS and Vercel, and how to prevent 502s reaching users.
• ImprovedBlog landing page on mobile now shows the latest two posts above the category filter so real content is above the fold.
• ImprovedHeader "Start Monitoring" button on free tools and comparison pages now meets the 44px touch-target minimum on phones.
• ImprovedHomepage caching policy is now applied cleanly at the edge, so return visits benefit from the stale-while-revalidate policy added in 2.19.1.
• FixProduction deployments now refresh the background worker container reliably when worker code changes.

An internal deploy fix. The previous release (2.19.2) had all the features (five new live-status pages, better content, performance and design) built and staged, but did not reach production due to an internal release-pipeline issue. This release ensures everything that was waiting now ships. No functional difference from 2.19.2; this is the version that actually reaches you.

• FixThe 2.19.2 release payload now deploys to production. An internal release-pipeline issue had held it back.

Added Venmo, Apple Pay, Apple Music, Apple TV+, and Rockstar Games (GTA Online servers) to the live-status page set after keyword research confirmed high demand and low competition in their SERPs. Total live-status pages now at 79.

• NewNew live status page at /tools/status/venmo for Venmo payments, sign-in, and bank transfers.
• NewNew live status page at /tools/status/apple-pay for Apple Pay in-store, checkout, Apple Cash and Express Transit.
• NewNew live status page at /tools/status/apple-music for Apple Music streaming, library sync, and iCloud Music Library.
• NewNew live status page at /tools/status/apple-tv for Apple TV+ streaming, MLS Season Pass, and live events.
• NewNew live status page at /tools/status/rockstar for Rockstar Games servers including GTA Online and Red Dead Online.

A focused pass following a full SEO audit. Every live-status page now carries genuinely unique per-service content (not just a template). Mobile pages load and render better. Marketing copy is honest about what Visual Sentinel actually is (built by a DevOps engineer, not a generic full-stack developer). Compare pages no longer confusingly imply that we sell our competitors.

• NewEvery one of the 74 live-status pages at /tools/status/[service] now carries a unique "About" paragraph and 2 service-specific FAQ answers, written by hand, instead of sharing one template.
• ImprovedStatus pages stopped jumping around at load. The live status card now renders in a reserved height so there is no visible shift when the check completes.
• ImprovedHomepage hero title resized on phones so it no longer wraps to 6 lines and hides the value prop below the fold.
• ImprovedPaddle checkout script no longer loads on pages that do not need it (homepage, blog, tools, compare). About a second faster first paint on those pages.
• ImprovedImages site-wide now serve AVIF first, smaller and sharper than WebP on modern browsers.
• ImprovedBrowser back-button navigation is now instant again. Previous cache-control settings were forcing a full re-render on every back press.
• FixAbout page now correctly describes the founder as a DevOps engineer and sysadmin, not a generic full-stack developer.
• FixBlog post meta author tag now matches the schema (Acefina LLC) on every post.
• FixCompetitor comparison page titles fixed a stray space before the comma that read as a typo.
• FixPricing page title shortened to "Simple, Transparent Pricing" so the keyword sits at the front.
• Fix/blog/<missing-slug> 404 page no longer renders the header twice.

We scaled the free live status checkers from 27 services to 74 with AI platforms (ChatGPT, Claude, Gemini, Perplexity, Midjourney), consumer tier 1 (YouTube, Netflix, Instagram, Facebook, Twitter, Gmail, TikTok, Roblox), and cloud / dev infrastructure (Cloudflare, AWS, Azure, Google Cloud, Zoom, Slack, Dropbox, Heroku, Netlify, DigitalOcean, Jira, GitLab, Firebase, and more). Every status page ships with real-time checks from our EU and US nodes, a proper FAQ block, and a one-click "Monitor this yourself" CTA. Blog URLs are now year-free so the same post can be refreshed across years without breaking any inbound links.

• NewNew live status pages for ChatGPT, Claude, Gemini, Perplexity, Midjourney, GitHub Copilot, and Cursor. Check if any of them are down right now at visualsentinel.com/tools/status/<service>.
• NewNew live status pages for YouTube, Netflix, Instagram, Facebook, Twitter, Gmail, Amazon, TikTok, Snapchat, Roblox, Xbox, PlayStation, Minecraft, Twitch, Steam, PayPal, Outlook, Hulu, Pinterest.
• NewNew live status pages for Cloudflare, AWS, Google Cloud, Azure, Zoom, Slack, Dropbox, iCloud, Disney+, Max, Heroku, Notion, Netlify, DigitalOcean, Jira, Zendesk, Trello, Telegram, GitLab, Firebase.
• ImprovedBlog URLs are now year-free. Same post, refreshable across years, no broken backlinks.
• ImprovedDuplicate blog posts consolidated into single canonical URLs so the best version of each guide is the one that ranks.

A focused release on making the product tell the truth. Monitors with DNS problems now correctly show as down (they were being silently marked operational). Cloudflare imports skip zones with no origin. Competitor comparison tables were rebalanced to credit the competition fairly. New /cookies, /accessibility, and /bot pages published. Em-dashes removed from user-facing copy per house style.

• FixMonitors whose domain has no DNS records (parked, freshly-imported Cloudflare zones, expired domains) now correctly show as "Down: DNS resolution failed" instead of "Operational, 100% uptime, 0 ms".
• FixCloudflare import now pre-checks each zone for an A or AAAA record and skips zones with no origin, surfacing them in the import summary with a specific notice so you can fix DNS in Cloudflare before retrying.
• NewMonitor detail page now shows a clear amber warning if the monitor has no alert channels attached, with one-click access to the settings tab.
• FixSetup-progress gauge now counts email as a valid alert channel. Accounts that used only email were previously stuck below 100% indefinitely.
• FixMonitor list headers now break down the status counts (X operational, Y down, Z degraded, W paused) instead of showing only the operational number.
• NewNew /cookies page lists every first- and third-party tracker we use, why we use it, and how to manage your preferences.
• NewNew /accessibility page stating our WCAG 2.1 AA commitment, known limitations, and how to report issues.
• NewNew /bot page with our stable monitoring User-Agent and ready-to-paste WAF/firewall rules for Cloudflare, AWS, and Nginx so customers can whitelist our checks reliably.
• ImprovedCompetitor comparison pages rebalanced. The UptimeRobot table now credits UptimeRobot for keyword monitoring, port monitoring, and SMS/voice alerts that were previously omitted.
• ImprovedNotification channel count reconciled to 11 across every mention on the site (homepage, plan cards, FAQ, notification-channels page).
• ImprovedRegion claims corrected from "3 global regions (EU, US, Asia)" to "2 regions (EU + US)" across the site. Asia coming soon.
• ImprovedLimited-time discount banner renamed to "Launch Offer" with the real combined discount percentage shown explicitly, so the banner does not contradict the actual checkout price.
• SecurityClickjacking protection restored sitewide. Public status pages remain embeddable; the dashboard and marketing site no longer are.
• SecurityPublished a dedicated security contact at /.well-known/security.txt for responsible disclosure.
• FixWorker no longer rebuilds every monitor scheduler on every restart. Reduces the window where monitors could go dark during a worker crash.
• FixAPI 404 responses are now JSON for API routes instead of a 25 KB HTML page. Clients integrating with our API will no longer crash on JSON parse errors.

A sweep of marketing copy, trial-state emails, and conversion UX. Stats and reviews now show only what is real. New monitors are live immediately (instead of waiting for a worker restart). Free plan users no longer see a paywall.

• ImprovedPublic stats and review counts now reflect reality. Any inflated numbers and unverified rating schema have been removed.
• ImprovedTrial expiry now keeps you on the Free plan with your 3 most recent monitors running, without any paywall. Existing trial emails have been rewritten to match.
• FixNew monitors now receive recurring checks immediately after creation (previously only an initial check fired, and recurring checks started on the next worker restart).
• FixBroken dashboard links across the server alert banner, incident detail, AI insights, and monitor-form flows now resolve correctly.
• FixWelcome email now shows the correct monitor limits for Business (60) and Agency (300).
• Fix"Live chat" banner, "3 regions (EU, US, Asia)" claim, and the stale 14-day trial references across blog posts, the AI chat assistant, and API error messages have been corrected.
• SecurityMeta Pixel now fires only after explicit opt-in on EU traffic. Enhanced security hardening on a few mutation endpoints.

• FixThe Free plan CTA on the pricing page and billing settings was sending logged-in users to a 404. Now opens the dashboard overview as intended.
• FixThe "Welcome to the Free plan" email button was pointing at a non-existent /dashboard URL. Fixed.

The onboarding plan-selection page now shows the Free plan alongside paid tiers and uses the new single-story trial messaging.

• ImprovedPlan selection page now shows a Free plan card next to the paid tiers. Pick Free to start a 7-day full-feature trial with no card, or pick a paid plan for a 10-day trial with a card on file.
• FixRemoved the stale trial-mode toggle and old "card saved, not charged until day 15" copy that was left over from the earlier trial model.