RealUptime Blog
Notes on keeping infrastructure honest.
Writing from the RealUptime team on monitoring, incident response, and what we're building. New posts every few days.
Pingdom alternatives now that it is part of SolarWindsWhat Pingdom's pricing page publishes under the SolarWinds name, what a 25-check plan costs a year, and how to pick an alternative without guessing.Statuspage pricing: what subscriber counts cost youStatuspage prices public pages by subscriber tier, from free at 100 to $1,499 a month at 25,000. Here is each tier's yearly cost, with the arithmetic.GlitchTip vs Bugsink: two small Sentry-compatible trackers, and when neither is the answerGlitchTip is MIT-licensed and needs Postgres and workers; Bugsink is PolyForm Shield and runs in one container. License, pricing and hosting costs compared.Is Railway down? What Railway's own status page admits it will not show youRailway's status page says it reports only significant, widespread incidents. What that leaves out for your deploy and how to check your own app.Should you build your own status page? What DIY actually takesBuilding a status page takes a weekend. Making it trustworthy takes weeks: self-updating checks, subscriber email, components, custom domains and history.Rollbar pricing, explained: what the sliding scale and AI credits actually costRollbar Essentials costs $9 to $6,499 a month depending on event volume, plus metered AI credits. What that means at real volumes, against a flat price.Monitor a Coolify app from outside the serverCoolify health checks confirm a container answers on localhost, not that your app is reachable. Where they fall short and how to check from outside the box.RealUptime is now part of TreepaiOn October 6, 2026, Treepai, Inc. became the owner of RealUptime LLC. RealUptime continues as its own company with the same four products, prices and support.Status page subscriber notifications: what to send, when, and how not to become noiseWhich channels to offer status page subscribers, which events should send a notification, and how to keep the list from training people to ignore you.How to tell which deploy caused an error (and what to stop claiming when you can't be sure)How release tagging shows which deploy caused an error, what a suspect-commit feature can honestly claim, and a ten-minute setup for your CI.External uptime monitoring for a Hetzner server: why SysMon is not enough on its ownHetzner SysMon checks ping, ports and HTTP from Hetzner's own network and proves only a status code. What it misses and how to add an outside check.Free status page tools in 2026: what "free" actually includesInstatus, Statuspage, Better Stack, OpenStatus and RealUptime free status page plans compared on subscribers, custom domain and monitors, from pricing pages.Why your error tracker keeps grouping unrelated bugs into one issueWhy an error tracker puts different bugs in one issue, how Sentry documents its grouping, and the fix: normalize messages, drop line numbers, hash in-app frames.Is Supabase down, or is it your project? How to tell in under a minuteSupabase's status page lists components per region but shows only what Supabase publishes. How to tell a platform incident from a problem in your project.Status page vs uptime monitor: what each one actually does, and do you need bothAn uptime monitor detects outages and a status page tells customers about them. What each does, where they overlap, and when you need one or both.Website down from some locations but not others: what causes it, and how to prove itWhy a site can be down in one country and up in another: five causes, a triage order to find yours, and why one monitoring location cannot show it.Maintenance windows: how to take a service down on purpose without paging anyone or lying about itA maintenance window should silence operator alerts, never the measurement or public status. Schedule one so subscribers hear first and overruns still alert.Downdetector says down. The status page says fine. Why "is it down" sites disagree, and which one to believeDowndetector, IsDown, StatusGator and probe checkers use four kinds of evidence. What each can and cannot know, and how to read them together.Is the internet down, or is it just you? How a provider incident is actually measuredWhen five apps fail at once, is it Cloudflare, an AWS region or your ISP? A two-minute triage, and how RealUptime names a provider incident from probe readings.Sentry's free tier: what 5,000 errors a month gets you, and what happens afterSentry's free Developer plan allows 5,000 errors a month and one user. What happens past the quota, what the $26 Team plan adds, and a 10,000-event alternative.Certificates expire on a schedule. The outages are optional.Expired TLS certificates took down LinkedIn, Microsoft Teams and a mobile network serving 32 million people. How to get alerted 30 days before expiry.What a TCP port check proves, and what it does notA TCP port check shows whether a host accepts a connection on a port. When to use it for databases, mail and SSH, and when an HTTP check fits better.Your cron job stopped and nobody noticed. Heartbeat monitoring, explained.A heartbeat monitor alerts when a cron job's ping goes silent. How it works, how to size the grace period, and the three mistakes that make it useless.How often should an uptime monitor check your site? Less than the interval decidesA monitor's check interval only sets a ceiling on detection time. Confirmation, timeouts and region quorum decide the real number. The math, with examples.ChatGPT is down. What the status page said, minute by minuteIs ChatGPT down right now? A timeline of OpenAI's own banner and incident feed on August 31, and why a status page is only the vendor's claim.What it takes to track services honestly195 active services as of 2026-10-05. The count matters less than the bar a service must clear and how we label what we cannot measure.RealUptime is publicWe launched on Product Hunt, AlternativeTo and SaaSHub. Here is what Status, Monitor, Errors and Outages do, the free limits, and a 25% launch code.Uptime monitor false positives: the page you should not have gottenHow RealUptime decides a failure is real before paging anyone: a 3-minute region rule, regional disagreement, and our own errors kept off your record.When the monitor is the problemHow an uptime monitor fails silently, and what RealUptime does about it: region locks, drifting intervals, errors on our side, and who gets told.One internet, four speedsWe time the same public services from ten regions. Tokyo's median response is about twice Frankfurt's, and the gap changes what an alert threshold should be.Status pages that can't lieHow RealUptime status pages stay honest: per-region checks, templated incident text, one set of numbers, and provider claims labeled as claims.