Glossary
Degraded performance
Degraded performance means a service is still up but not working properly: responses are slow, some requests fail, or a feature is unavailable.
Also called Service degradation, Partially degraded service, Partial outage, Brownout.
Degraded performance, also called service degradation, is the state between up and down: the service still responds, but slowly, intermittently, or with some feature missing. It is what a status page means by "degraded" or "partially degraded service": most things work and some users are having a bad time. None of it is an outage, and for the people affected the distinction is academic.
Typical examples: search still works but takes eight seconds. Uploads succeed for small files and time out for large ones. One database shard is unavailable and a fifth of customers see errors. Checkout works for every payment method except one.
Partially degraded service is degraded performance with a boundary: the problem is confined to one region, one feature or a share of requests, and the rest of the service works normally. It is the label for a page where most users are unaffected and some are not. At the far end, a feature that has stopped working altogether is usually reported as a partial outage.
It is harder to detect than an outage precisely because the naive check passes. A monitor asking only "did it respond" reports a healthy service throughout. Catching degradation requires thresholds on response time, checks that assert on content rather than status codes, and monitors on individual features rather than on the front door alone.
It deserves a distinct state rather than being rounded to up or down. Rounded up, the status page contradicts every customer currently having a bad time — the fastest way to lose trust in a status page. Rounded down, the announcement of a total outage is untrue in the other direction and triggers a response the situation does not need. A separate degraded state is what lets the page say the thing that is actually happening.
It is also frequently the early phase of an outage. Latency climbing towards the timeout, error rates rising from nought to a few percent — alerting at the degraded stage buys the time that turns a potential outage into a paragraph in a weekly review.
| Status | What it means | Example | How a monitor catches it |
|---|---|---|---|
| Operational | Everything works at its normal speed | Pages load in their usual time and every request succeeds | Every check passes within its response-time limit |
| Degraded performance | Up, but slower or less reliable than normal | Search takes eight seconds instead of one | A response-time threshold, or a content check that fails while the status code is still 200 |
| Partial outage | One part is down for some users | Large uploads fail, or one feature returns errors | A separate monitor on each feature, not only the home page |
| Major outage | The service does not work for most or all users | The site returns 502 or does not answer at all | Any check: the request fails or times out |
Questions
What does degraded performance mean?
It means a service is working, but not properly, which is also called performance degradation or service degradation. Pages load slowly, a share of requests fail, or one feature is unavailable while the rest works. It is neither an outage nor normal operation.
What is a partially degraded service?
A service where only part of it is affected: one region, one feature, or a fraction of requests. Status pages use the phrase to say that some users are having problems while most are not.
What is the difference between degraded performance and an outage?
In an outage the service does not answer at all. When it is degraded it still answers, but too slowly or with errors for some requests, so a check that only asks "did it respond" can pass the whole time.
What is the difference between degraded performance and a partial outage?
Degraded performance means everything still works, just slower or less reliably than normal. A partial outage means one part has stopped working for some users, such as a feature or a share of requests that fail outright. Status pages that show both treat a partial outage as the more serious of the two.
How do you detect degraded performance?
Put a threshold on response time, check the content of the response rather than only its status code, and monitor individual features rather than the home page alone. In Uptimely, a monitor's criteria can set it to Degraded rather than Offline when a response-time or content rule matches.
In Uptimely
Uptimely carries degraded as a first-class monitor status, distinct from operational and offline, with its own status-page colour and its own notification rules. Monitor criteria can raise it on response time or content rather than only on a failed request.
Uptimely tracks degraded as its own monitor status, next to operational and offline, and shows it on your status page.
Start monitoring free →Read next
Related terms
Part of the Uptimely monitoring glossary — plain definitions of the vocabulary uptime monitoring, incident response and status pages are discussed in. Something here read wrong to you? Email support@getuptimely.com.