Response time monitoring: slow is the outage that comes first
Most outages start as a slowdown: a query that gets heavier every week, a disk that fills, a memory leak. StatusEagle records how long every check takes, from every location, so you can see the trend before it turns red.
Example check · demo data
Response time monitoring measures how long a website, API or server takes to respond to each check, and tracks those measurements over time and by location to reveal slowdowns and performance regressions.
What a synthetic response time includes
For an HTTP check, the time runs from sending the request to receiving the response: DNS lookup, connection, TLS handshake, your server’s processing time and the transfer. It does not include loading images and scripts or rendering in a browser. That’s what Core Web Vitals from real users measure.
Read medians and outliers, not averages
An average blends good and terrible into one misleading number. If 58 of 60 checks take 200 ms and two take 9 seconds, the average is about 490 ms: “a bit slow”. The truth is “fast, with two awful moments”. StatusEagle shows the typical value and the slowest checks separately, per location.
| Pattern | Likely cause | Gradual rise over days or weeks | Sudden step after a deploy | Spikes at the same time each day | One location much slower | Slow first check after a quiet period |
|---|---|---|---|---|---|---|
| Growing table without an index, cache hit rate falling, log volume | ||||||
| A regression in that release | ||||||
| Cron jobs, backups or batch imports competing for resources | ||||||
| Distance or routing, not your server | ||||||
| Cold starts (serverless) or an evicted cache |
Set a response-time budget
Pick a number per endpoint that reflects what users will tolerate: say 500 ms for an API call and 1 second for a server-rendered page. Anything consistently above it is a performance bug, even if it’s “up”. Confirm whether response-time threshold alerts are available per monitor and document the setting
Getting clean data
Related monitoring
Pair these monitoring checks to ensure complete stack coverage.
Frequently asked questions
Still deciding? Start monitoring free and verify your first service in ~2 minutes.
Measuring how long a service takes to answer each check, and tracking it over time and by location.
Your first monitor takes about two minutes.
Free plan: 10 monitors, 3 locations, 1 status page. No credit card.