Response Time Analytics
Response time is a leading indicator of performance problems. StackBloom Monitor records the time it takes your server to respond to every check and surfaces trends, outliers, and slowdowns so you can act before users start complaining.
Step 1: Open a monitor and click the Analytics tab
From the Monitor dashboard, click on any HTTP or API monitor to open its detail view. Select the Analytics tab at the top of the page. The tab is available for all monitor types that measure response time, including HTTP, API, and Page Speed monitors.
- Response time data is collected on every check interval — as frequently as every 30 seconds on Pro
- Analytics are available from the day the monitor was created; historical data is retained permanently
- Multiple monitors can be compared by opening each in a separate browser tab
- The Overview tab shows the current uptime summary; Analytics shows the performance trends
Step 2: View average, p95, and p99 response time metrics
The analytics chart plots three response time series simultaneously so you can see both typical and worst-case performance. Each metric tells a different story about your server's behavior.
- Average — the mean response time across all checks in the period; useful for capacity planning
- p95 — 95% of checks completed within this time; a good proxy for "most users' experience"
- p99 — 99% of checks completed within this time; reveals the worst-case slowdowns affecting a minority of requests
- Toggle any series on or off by clicking its label in the chart legend
Step 3: Set a response time alert threshold
Click Add Alert Rule from the Analytics tab and chooseResponse Time Exceeds. Enter a threshold in milliseconds. Monitor will fire an alert whenever a check result exceeds that value, letting you catch slowdowns as they happen.
- A good starting threshold is 2× your average response time (e.g., if average is 300 ms, alert at 600 ms)
- Set a consecutive-failures count of 2–3 to avoid alerts from single isolated slow checks
- Alert notifications go to all channels configured on the monitor (email, Slack, SMS, webhook)
- Response time alerts auto-resolve and send a recovery notification when times drop back below the threshold
Step 4: Filter by time range to spot trends
Use the time range selector at the top right of the Analytics tab to zoom in on specific periods. Narrowing the range to a single day shows hourly patterns; expanding to 90 days reveals longer seasonal or growth-related trends.
- Preset ranges: Last 24 hours, Last 7 days, Last 30 days, Last 90 days, Custom
- Hover over any data point on the chart to see the exact timestamp and response time value
- Look for daily patterns — many sites are slowest during peak business hours or nightly batch jobs
- Compare two periods by opening the same monitor in a second browser tab with a different date range
Step 5: Export response time data as CSV
Click Export > CSV from the Analytics tab to download a time-series file of every check result within the selected date range. Each row includes the check timestamp, response time in milliseconds, HTTP status code, and whether the check was considered up or down.
- CSV exports are scoped to the currently selected date range and monitor
- Import the CSV into Excel, Google Sheets, or a BI tool like Looker for custom reporting
- Large exports (90+ days of frequent checks) are delivered to your email as a download link
- Use the Monitor API to automate regular data pulls without visiting the dashboard
💡 Tip: A sudden spike in p99 response time — while average stays normal — often indicates a backend bottleneck affecting a subset of requests, such as a slow database query, a memory leak under load, or a third-party API timing out. Investigate p99 spikes early before they broaden into average slowdowns that every user feels.