Oh Dear vs UptimeRobot

Oh Dear

Failure confirmation is documented properly: each monitor has a primary location, a failure there is re-checked from a different city on a different provider, and both must report down before a downtime notification goes out — with the probe addresses published and updated automatically. HTTP checks carry keyword, forbidden-string, header and JSON body assertions, alongside TCP, ping, DNS, domain expiry, SSL chain validation and cron heartbeats, from sixteen named cities at a shortest documented cadence of one minute. The vendor itself states the checks are strictly outside-in with no agent and that the TCP check is not a synthetic checkout, and we found no public information on multi-step or browser transaction checks.

UptimeRobot

Nine monitor types plus heartbeat, SSL and domain expiry, ICMP with traceroute and response times, JSON assertions with all-or-any pass logic, DNS checks, maintenance windows and a 15-second interval are documented, and a failed check is re-checked from a different data center in the same region before an alert goes out — the confirmation logic I always ask for, actually written down. The re-check stays within one region and regions are selectable without a city-by-city list, and we found no public information on browser or transaction checks, private locations inside the customer's network, or published probe health.

Oh Dear

Twelve channels including SMS through the customer's own Vonage account, PagerDuty and Opsgenie with automatic resolve, acknowledgement with fixed durations and hourly reminders, recovery notifications, maintenance periods, throttled repeats, and per-destination event filtering with business-hours routing — the noise-reduction end is genuinely documented. We found no public information on phone call alerts, on-call schedules and rotations, or escalation chains with timeouts, and SMS costs sit with the customer's own Vonage account rather than on the vendor's price list.

UptimeRobot

Every channel I'd ask about is documented — email, SMS, phone call, mobile push, Slack, Teams, webhooks, and forwarding to PagerDuty — with per-monitor contact assignment, an up-to-60-second timeout to prevent false positives, and a 3-of-5 consecutive-request rule before a response-time alert fires. SMS and voice costs are published in credit packs with voice calls flat at two credits and a country-by-country credit list, and the pages even disclose that a failed auto-refill charge switches auto-refill off with no notification sent — an honest detail I respect. Escalation policies and on-call schedules, though, are only reachable through third-party integrations like ilert and Zenduty, and we found no public information on native escalation with timeouts, acknowledge-from-app or chat, alert grouping, or recovery notifications.

Oh Dear

Custom domain with a certificate the vendor issues, connected checks feeding the page automatically (with a toggle so a person can approve every post), updates with severities including Scheduled stacking into a timeline, email, Slack and RSS subscribers, twenty-five languages, saved templates and an embeddable badge — unlimited pages on every plan. Access is public or behind a secret URL, which the vendor itself says is not a login and not per-user access control. The pages are documented as hosted separately from the monitored site, but we found no public information on separation from the vendor's own service, component groups, or per-component and SMS subscriptions.

UptimeRobot

Custom domain with automatic Let's Encrypt certificates, password protection, eleven languages, whitelabeling, embeddable badges, incident and maintenance announcements and email subscribers are all documented, with per-plan page counts from one to unlimited. Two caveats the pages themselves state: the custom domain must be a subdomain — apex domains are unsupported — and a downgrade, expired trial or failed payment immediately blocks the whole page on the custom domain. We found no public information on SMS, RSS or webhook subscribers, component groups, incident templates, or how the status page hosting is kept separate from the main service.

Oh Dear

A documented REST API with bearer-token authentication covers monitors and status page updates, a MIT-licensed Terraform provider (maintained by a third party) and a JSON-template workflow keep check definitions in version control, and SAML single sign-on is on every plan and can be required for everyone but the team owner. Unlimited users with roles, two-factor, tags and tag groups, client grouping, guest access limited to selected monitors and status pages, and a reseller program round out the team side. We found no public information on SCIM provisioning, audit logs or bulk import, and the vendor states its reports are not fully white-labelled.

UptimeRobot

The v3 REST API covers monitors, alert contacts, maintenance windows and public status pages with documented rate limits, OAuth 2.0, key rotation and bulk operations, alongside an official CLI and MCP, read/write/notify-only team roles, and two-factor authentication required for privileged admin access per the security annex. Terraform and Ansible appear as community tooling for onboarding monitors via infrastructure-as-code rather than a documented first-class provider. We found no public information on SAML single sign-on, SCIM provisioning, audit logs, or client sub-accounts with their own logins.

Oh Dear

Retention is stated in plain words — monitoring data typically one to two years per plan, certificate history kept, every past monthly report kept on a browser link — export of all monitoring data is offered via the API, probe addresses are published in JSON and CSV and updated automatically, and response times break down into DNS, connect, TLS and time to first byte. Against that, the vendor itself states no formal SLA with credits is offered, only a 99.9% target, and we found no public information on the vendor's own status page, its incident history, post-mortems or response-time percentiles.

UptimeRobot

The reporting engine is real: uptime percentages for 24 hours to 365 days, MTBF per monitor and account over 7, 30 or 365 days, monthly uptime reports with downtime budgets, SLA policies with 99.9% and 99.99% targets whose uptime is the average of member monitors shown to three decimals, per-client policies each with their own report, and retention stated per plan at 3 to 24 months. But the questions I ask of the monitor itself go unanswered: we found no public information on the vendor's own public status page or incident history, an SLA with credits, published probe IP addresses for allow-listing, response time percentiles or breakdowns, or a raw export of check results.

Oh Dear

Immutable VOF in Berlaar, Belgium is the contracting entity under Belgian law, primary data sits with Combell in Belgium, a data processing agreement with the EU's 2021 standard contractual clauses and a thirty-day subprocessor objection window is published, and the subprocessor list is public with each processor's location and certifications. The vendor is frank that probe nodes, backups and check reports run on US providers, that it claims no CLOUD Act immunity, and that the ISO 27001 certificate belongs to the host rather than to Oh Dear itself. That honesty earns the level, but backups and check reports on US providers and no documented self-hosted edition keep it short of sovereign end to end.

UptimeRobot

The contracting entity is UptimeRobot s. r. o. of Bratislava with Slovak governing law and Slovak courts in both the terms and the DPA, a published DPA version 2.2 with SCC modules and a docking clause, 30-day subprocessor change notice, full liability for subprocessors, deletion within 90 days with written certification, backup deletion within 180, and breach notice within 72 hours. Subprocessors are named with purposes — Hetzner hosts the monitoring infrastructure, API services and status pages — and processing is stated as primary European Union with a United States secondary under stated safeguards, with encryption in transit and at rest. It stays below the top because UptimeRobot itself holds SOC 2 Type I while the Type II and ISO 27001 certifications belong to its infrastructure providers, and individual subprocessor locations are not stated.

Oh Dear

Nine tiers from Solo at €15/mo for two sites to Scale at €399/mo for two hundred are public, and the only dimension that scales is the number of monitored sites — every feature, unlimited users, single sign-on and unlimited status pages sit on every plan, billed monthly or annually, cancel anytime, with a thirty-day money-back guarantee and thirty days' notice of price increases. SMS runs through the customer's own Vonage account so no per-message price appears on these pages, no overage handling is described, Enterprise at 200+ sites is custom, and the captured pages give different figures for VAT — the pricing page says EUR ex. VAT while the terms say prices include applicable Belgian VAT where required.

UptimeRobot

The free tier is fully spelled out — 50 monitors, 5-minute interval, one status page, no credit card — and paid tiers publish interval, status-page count, seats, retention and included SMS credits, with extra seats at $15/19 a month, SMS and voice credit packs from $3 for 10 credits, voice calls flat at two credits, and a country-by-country credit consumption list in a blog post. The pages even disclose the ugly edge: a failed auto-refill charge switches auto-refill off with no notification sent, and auto-refill needs a card, excludes PayPal and the smallest pack. But Solo and Scale are listed as starting prices, extra monitors and shorter intervals are sold as Upgrades without a captured unit price, and we found no public information on VAT treatment or monthly list prices.