Project Monitoring
From the Project sidebar, open Monitoring to view metrics for all containers within the Project. If the Project contains multiple containers, their metrics are displayed together on the same chart with separate legends for comparison.
Filters and Display Options
Container Monitoring
From an individual container page, open Monitoring to view detailed metrics for that container. Use the metric selector in the upper-right corner to customize the dashboard and display only the metrics you need.
Available Metrics
Replica Timeline
The Healthy Replicas Count card at the top of the container’s monitoring page shows when Autoscaling added or removed Replicas, alongside your other metrics.
- Replica count and runtime: The line shows how the Replica count changed over the selected range, and the rows below it show when each Replica was running. Hover a bar to see the Replica ID and when it started and stopped.
- Slots: Each row is a slot that Replicas take turns using, not a single Replica. When a Replica stops and a new one starts, the new one reuses the free row, so the number of rows equals the most Replicas that ran at the same time.
- Slot labels: A row used by only one Replica shows its ID. A row used by several shows Slot N, with ×N on the right for how many Replicas ran in it. If there are more than 8 slots, click Show N more slots to see the rest.
- Summary: The line above the rows shows how many Replicas ran in the range, how many are still running or have terminated, and the median lifetime of the terminated ones.
- Short-lived Replicas: A Replica that ran only briefly appears as a dot. Dots that sit close together are grouped into a striped band.
- Min and max guides: The Autoscaling minimum and maximum Replica counts are drawn as dashed lines, so you can see how close you are to the limit. If Autoscaling is off, the card shows the fixed Replica count instead.
- How Replicas are scaled: Hover the ⓘ icon on the card to see the scaling rule. With Autoscaling on, a Replica is added when each Replica is handling more than the target number of concurrent requests on average, and removed when the load drops. With Autoscaling off, a change in the Replica count may mean a restart or a failure.
- Scale markers: Every metric chart marks the moments the Replica count changed with a vertical dashed line, so you can compare how requests, latency, GPU utilization, and other metrics changed before and after each scale event.
Autoscaling decides only from the number of concurrent requests per Replica. The scale markers help you compare metrics before and after a scale event; they do not mean that the metric on that chart caused it.
Expanding a Band
Click a striped band to list the Replicas grouped in it.
- The header shows the band’s time range, how many Replicas it holds, and their median lifetime.
- Each row shows a Replica’s ID, start and end time, and lifetime. Hover a row to mark that Replica on the band.
- Click Zoom into this range to narrow every chart on the page to the band, with a little margin on each side. Click Reset zoom to go back.
Replica List
Click List in the upper-right corner of the rows (outlined in red below) to open a table of every Replica in the range, with its status, start and end time, and lifetime. Once the table is open, the button reads Hide list; click it again to close the table.
- Search by Replica ID, and sort by Most recent or Longest lived.
- The table shows 10 Replicas per page. Use the arrows in the lower-right corner to move between pages.
- Hover a row to highlight that Replica in the timeline above. This is useful for telling apart the Replicas that ran in the same slot (a row marked ×N).
Drag to Zoom
Drag across any chart to zoom every chart on the page into that range. The data is re-queried at a finer interval for the zoomed range. Click Reset zoom to return to the previous range.Container Runtime Logs
From the Runtime Logs page of a container, you can view logs for both running and terminated Replicas.
- Logs from both running and stopped Replicas remain viewable.
- Download saves the selected Replica’s logs, and Download all saves every Replica’s logs at once.
Live Logs
Logs from running Replicas are appended to the bottom of the view as they arrive, with no refresh needed. They usually appear within a few seconds.- Follow latest logs: On by default, so the view scrolls to the bottom with every new line. Turn it off and your position stays put; a new button appears at the bottom and jumps to the newest line.
- Pause live updates: Click the Live button in the log view to stop showing new lines for a while. Click it again to load what you missed and resume.
- If the connection drops: The view reconnects automatically and fills in the lines from while it was disconnected. A browser tab left in the background for more than 5 minutes disconnects, then fills in the gap when you come back.
- Stopped Replicas: Show their past logs only, with no live updates.
Long Logs
- The newest 5,000 lines appear first, and older lines load in the background 10,000 at a time and are added above. Progress is shown at the top while they load.
- The view holds up to 50,000 lines. Older lines are not loaded, and search only looks through the loaded lines.
- Download saves the loaded lines with the current filters applied. If you download while older lines are still loading, only the lines loaded so far are saved.
Hide Health Checks and Metrics
Turn on Hide Health Checks & Metrics under View Options to hide health-check and metrics-scrape request lines (GET …/health, GET …/metrics) and see only your application’s logs.
Log Alerts
Get notified when something shows up in the logs, or when logs stop. Create rules from the Log alerts button above the left panel, or from the bell icon in the search box. See Log Alerts for details.If a deployment fails or behaves unexpectedly, checking the runtime logs is usually the fastest way to identify the cause.
Related Documentation
Troubleshooting
Learn how to diagnose common deployment and operational issues.
Autoscaling and Scheduled Scaling
Learn how Replica counts change with Autoscaling and Scheduled Scaling.

