The term covers two situations. If you run a SaaS product, it means monitoring your own multi-tenant application the way customers experience it: per customer, per region, and against the uptime and response-time promises in your SLAs. If your business depends on SaaS tools and APIs, such as payments, identity, email delivery or a CRM, it means watching those services too, because their outages become your outages even though you can't see inside them.
The phrase is also used for monitoring tools that are delivered as SaaS rather than self-hosted. That is a different question; see SaaS vs. self-hosted observability.
Monitoring the SaaS product you run
- Availability and response time against your SLAs and internal service level objectives.
- A per-tenant view of errors and latency. Averages across all customers hide the one large customer having a bad day, or the tenant whose usage is slowing everyone else down.
- Real user experience: page load times, Core Web Vitals and JavaScript errors in customers' browsers.
- API performance and usage by customer and endpoint, if customers integrate with your API.
- Background work: queues, scheduled jobs and webhooks, which fail quietly.
- Infrastructure and cost per service, so growth doesn't outrun margins.
Monitoring the SaaS you depend on
- Measure from your side. Track the latency and error rate of every outbound call your application makes to a third-party API.
- Run your own checks against the endpoints you rely on, such as a login provider or a payment API, so you find out before your users do.
- Don't rely on vendor status pages alone. They describe the vendor's whole service, not your account or region, and they are usually updated by hand after the fact.
- Keep evidence. Your own data is what you need to hold a vendor to its SLA.
Metrics worth tracking
- Uptime percentage per service and per region
- Latency percentiles (p95, p99) per endpoint and per tenant
- Error rate, split by your errors and third-party errors
- Apdex or another user satisfaction score
- Usage per tenant, to spot abuse, churn risk or capacity needs
SaaS monitoring with Atatus
Atatus covers both sides. APM records external calls, so you can see the latency and errors of each third-party API your app uses, and produces SLA reports of requests, failures and Apdex by day, week or month. API analytics breaks API traffic down by customer, real user monitoring shows the experience in the browser, and uptime monitoring checks your endpoints, and the ones you rely on, with HTTP, SSL, DNS and multistep API checks.