A reliable website uptime monitoring process does more than send an alert when a page stops loading. It helps you confirm the failure, identify whether DNS, TLS, the application, the server, or the hosting provider is responsible, preserve useful evidence, and escalate the incident with enough detail to reach a resolution. This runbook provides a repeatable website outage response plan, including monitoring checks, alert thresholds, incident logs, hosting escalation steps, and a schedule for reviewing your setup.
Overview
Uptime monitoring is the practice of testing whether a website or service is reachable and behaving as expected. A basic monitor may request a homepage and report whether it receives a successful response. A stronger setup checks several layers separately: domain resolution, the HTTP response, the TLS certificate, response time, and an important application transaction such as a login or checkout flow.
No single check proves that a website is fully healthy. A homepage can return a successful status while an API, database connection, contact form, or authenticated area is unavailable. Conversely, one monitoring location may report a failure caused by a local network path rather than by your hosting environment. Treat each alert as a signal to verify, not as a complete diagnosis.
The purpose of this runbook is to make response consistent. Before an incident occurs, document the website’s normal behavior, monitoring contacts, hosting account details, DNS provider, domain registrar, and escalation path. Keep the runbook accessible even if the primary website is unavailable. A separate storage location, internal wiki, or downloadable copy can help during an outage.
What to track
1. Availability and HTTP behavior
Monitor the public URL over HTTPS and record the returned status code, response time, and final URL. Decide which responses are expected for your site. A normal page might return a successful 2xx response, while a protected page may intentionally return a redirect or an authentication response. Document those exceptions so the alerting system does not treat expected behavior as an outage.
Track repeated failures rather than isolated events. Useful fields include the first observed failure, the last successful check, the monitoring location, the status code, timeout type, and whether the result was a connection failure, DNS failure, TLS failure, timeout, or application error.
2. DNS resolution
Check that the domain resolves to the intended address or service. Review the relevant A, AAAA, CNAME, MX, and TXT records when the incident affects website access or email. Compare the result from more than one resolver if possible. A DNS problem may appear gradually because resolvers can hold records for different periods according to their cached values.
Keep a record of recent DNS changes, including the person who made the change, the time, the previous value, and the intended value. This is particularly important when connecting a domain to new web hosting, a content delivery service, or a third-party email platform. The domain and DNS setup guide can serve as a reference when checking those relationships.
3. TLS and certificate health
Use an HTTPS check that validates the certificate chain, hostname, and expiration rather than merely checking whether a port is open. Record certificate warnings, handshake errors, protocol failures, and unexpected redirects. Set an advance reminder for certificate renewal based on your operational process, and verify that automated renewal is actually completing successfully.
4. Application and performance signals
Add a synthetic check for a critical path, such as submitting a test form, loading a health endpoint, or completing a safe read-only transaction. Do not use a real customer action for testing unless the process is designed to prevent duplicate orders, messages, or other side effects.
Track response time separately from availability. A page that eventually loads may still be unusable, and a gradual increase in time to first byte can indicate resource pressure, an application regression, a database problem, or a change in traffic. For a broader performance workflow, see how to troubleshoot hosting speed, TTFB, and Core Web Vitals.
5. Evidence and ownership
Maintain an incident log with these fields:
- Incident ID, website, environment, and severity.
- First alert time, confirmation time, mitigation time, and recovery time.
- Monitoring locations, URLs, status codes, error messages, and screenshots.
- Recent deployments, DNS edits, certificate changes, plugin updates, or infrastructure work.
- Hosting account, domain registrar, DNS provider, and support case references.
- Current hypothesis, actions taken, owner, next checkpoint, and final cause.
Cadence and checkpoints
Choose check intervals according to the business impact of the website and the speed at which you need to respond. A low-risk informational site may use a less frequent external check, while a transaction or customer portal may require more frequent checks and a separate synthetic transaction monitor. The right interval should balance detection speed, monitoring cost, and the risk of false alerts.
Use a confirmation rule before paging someone for a single failed request. For example, require repeated failures within a short window, or confirmation from a second monitoring location. Do not delay escalation when the failure affects a critical transaction, produces a certificate error, or is confirmed by multiple independent checks.
Review the following checkpoints on a regular schedule:
- Each alert: Confirm the failure from another network or monitoring location, check the hosting status page, and record the initial evidence before changing configuration.
- After recovery: Confirm the homepage, key application paths, DNS resolution, TLS validation, and any affected email or API integrations.
- Monthly: Review alert noise, missed incidents, response times, certificate dates, backup visibility, and the accuracy of on-call contacts.
- Quarterly: Test access to the registrar, DNS provider, hosting control panel, monitoring account, and support channels. Verify that at least two authorized people can respond.
- After a change: Add a deployment or DNS checkpoint immediately after migrations, platform changes, major updates, and certificate configuration changes.
Keep domain and hosting ownership information current. A service may be online while its renewal, DNS, or administrative access is at risk. The domain portfolio management checklist helps organize ownership, renewals, DNS, and access reviews.
How to interpret changes
Start by separating the failure into layers. If the domain does not resolve, investigate DNS, registrar status, nameservers, recent record changes, and resolver differences. If DNS is correct but HTTPS fails, inspect the certificate, load balancer, firewall, origin availability, and redirect configuration. If HTTPS responds with a server or application error, review deployments, logs, resource usage, dependencies, and database connectivity.
A failure that occurs only from one monitoring location may indicate a regional network path, filtering rule, or monitor-specific problem. Confirm it from another location before declaring a global outage. A provider status page can show a known platform incident, but it should not replace checks against your own domain and application.
Escalate to the hosting provider when the issue appears to involve the platform, virtual machine, network, storage, managed database, load balancer, or another service outside your administrative control. Include the domain, affected service, UTC timestamps, monitoring locations, exact errors, recent changes, traceroute or request information when relevant, and the business impact. Ask focused questions: Is there an incident affecting the service? Can you confirm whether the host or network is healthy? Are there resource, firewall, or rate-limit events on the account? What evidence or next checkpoint does support require?
Do not make several unrelated changes at once. A rushed DNS edit, rollback, plugin deactivation, or cache purge can remove evidence and make the cause harder to isolate. Record each action and its result. If recovery requires a temporary workaround, document its owner and a removal date.
When to revisit
Revisit this runbook monthly or quarterly, and immediately after any outage, migration, provider change, domain change, major deployment, or monitoring-platform change. Update the URLs, expected status codes, certificate ownership, escalation contacts, hosting plan details, and support links. Remove checks for retired services and add checks for new critical paths.
After every meaningful incident, hold a short post-incident review. Confirm the timeline, distinguish the technical cause from the detection gap, and note whether the alert arrived early enough, reached the right person, and contained sufficient evidence. Then choose a small number of corrective actions, such as adding a DNS check, adjusting an overly sensitive threshold, testing a backup path, improving deployment rollback, or clarifying provider escalation instructions.
Use this closing checklist:
- Verify the website, critical paths, DNS, TLS, and email or API dependencies.
- Export or preserve monitor results, logs, screenshots, and provider correspondence.
- Complete the incident timeline and assign unresolved actions to named owners.
- Review whether alerts were too late, too frequent, or aimed at the wrong responders.
- Test administrative access and confirm that the next person on call can follow the runbook.
- Schedule the next monthly or quarterly reliability review.
A practical uptime program is maintained, not installed once. Regular review keeps monitoring aligned with the real website, while disciplined incident logs turn each outage into clearer evidence and a more dependable hosting escalation process.