Notes from the team
Architecture decisions, post-mortems, and pricing essays from the team behind ZnowPulse. New posts every couple of weeks.
Before adding more monitors, review the ones you already have
A compact monitoring review to find stale targets, misleading health checks, broken alert routes, and missing ownership.
Read postA status page is not an incident response plan
Make your status page useful by deciding who updates it, what evidence they need, and what happens when nobody is immediately available.
Read postBuilding a status page that survives its own outage
A reliable status page needs a different failure path from the systems it reports on. Here is the practical architecture to aim for and the trade-offs to document.
Read postWhat a hosted status page actually costs in 2026
A hosted status page has real operational costs. Here is what it needs to do during an incident, where pricing differs, and how to decide whether managed hosting is worth it.
Read postPlan maintenance without creating a monitoring blind spot
Separate customer communication, alert suppression, and health evidence when planning maintenance for a small service.
Read postThe monitor says down, but the site works for me. What now?
A calm checklist for investigating disagreements between your browser and an external uptime monitor.
Read post99.9% uptime: how much downtime is that, really?
Translate uptime percentages into minutes, then check the measurement window, sampling method, and exclusions behind the number.
Read postThe deployment passed. Now check the path customers use.
A practical post-deployment checklist covering public routing, read-only smoke tests, asynchronous work, and monitoring evidence.
Read postLiveness, readiness, and uptime are three different questions
Why a healthy container does not prove a reachable application, and how to separate restart checks from customer-facing monitoring.
Read postWhat to say during an outage: four status update templates
Clear, adaptable incident updates for investigation, mitigation, monitoring, and resolution—without invented ETAs.
Read postYour uptime alerts are noisy. Start with these four questions.
Reduce alert fatigue by separating symptoms, confirmation, routing, and recovery instead of muting everything.
Read postFive-minute vs ten-second checks: the detection math
Understand polling intervals, missed short outages, and confirmation delays before choosing faster uptime checks.
Read post200 OK, broken app: what your uptime check is missing
A successful HTTP response does not prove your application works. Learn which checks catch maintenance screens, broken dependencies, and empty responses.
Read postThe first five monitors a small SaaS actually needs
Start with customer-facing failures, not a wall of green dots. A practical monitoring plan for a small SaaS team.
Read post