Monitoring, Uptime and Responding to Incidents
What to monitor
- External HTTP checks from at least two regions, every minute, on a real page rather than a health stub.
- SSL certificate expiry, with alerts 30 and 7 days out.
- CPU, memory, disk space and disk I/O, with disk space alerting at 80%.
- Database connections and slow query counts.
- Backup success — a silent failed backup is the worst kind.
- Domain expiry, which takes down more sites annually than hardware ever does.
Alert thresholds that avoid fatigue
Alert on sustained conditions, not instantaneous spikes: CPU above 85% for five minutes, memory above 90% for five minutes, two consecutive failed HTTP checks. Route genuinely urgent alerts to phone notifications and everything else to email or chat, otherwise the team learns to ignore all of them.
Incident checklist
- 1 Confirm the outage from outside your own network before changing anything.
- 2 Identify the layer: DNS resolving, TCP connecting, TLS handshaking, HTTP responding, application rendering.
- 3 Check disk space and memory first — they cause a surprising share of mysterious failures.
- 4 Read the web server and application error logs for the failure window.
- 5 Apply the smallest safe fix, and record what you changed.
- 6 Restore from backup only when a fix is not identifiable quickly.
- 7 Write a short post-incident note: cause, fix, and the monitoring or change that prevents a repeat.
Our platform status
HostBotics publishes live service status so you can immediately tell whether an issue is platform-wide or specific to your site. Check it first during an incident — if the platform is healthy, the cause is almost always application-level.
Frequently asked questions
What uptime should I expect?
Our infrastructure targets 99.9% or better. Realistically, most website downtime originates in application code, plugin updates or expired certificates rather than server hardware.
How quickly are hardware issues handled?
Platform monitoring runs continuously and our team responds around the clock; you can also open a priority ticket for anything affecting production.
Need a hand applying this?
Our engineers can do it for you. Open a ticket and we will help.