Snapshot¶

In DeepCube Web, a snapshot shows how your grid was doing over the range selected in the time range picker, and which services sent traces in the 5 minutes before the end of that range. It answers the question a trace or a log line raises: what else was going on at that time?

A snapshot shows you two things:

• Four charts for the selected range: Apdex, request rate, error rate and p90. They are the same charts shown above the Traces and Logs lists.
• What was running: the services that sent traces in the 5 minutes before the snapshot time.

What Snapshot is for¶

Use Snapshot when a trace, a log line or a chart makes you ask what else was happening at that time. It helps you:

• Put a span or log line in context. From a trace or a log line, Around this span or Around this log lists the services that sent traces in the 5 minutes before it, with the span's or log's own service first.
• See the grid's overall shape for a time window. The four charts (Apdex, request rate, error rate and p90) cover the whole range you picked.
• Compare two periods. Change the range in the time range picker and the page follows, so you can look at a quiet period and a busy one with the same view.
• Share what you are looking at. A fixed range in the address holds the same start and end for whoever opens it. See Sharing a snapshot.

Snapshot reports what the grid received. It does not report the health, state or cause of any single service. Use the Traces and Logs lists for the detail behind a figure.

Open a snapshot¶

There are three ways in.

From the menu¶

Select Snapshot in the grid's navigation menu, directly under Logs. The page uses whatever range is selected in the time range picker, so you do not need to enter a time.

From a span or a log¶

1. Open a trace from Traces, or a log line from Logs.
2. In the Trace tabs, select Around this span (from a trace) or Around this log (from a log line).

The tab shows the grid at the time the span or log line was recorded, with that span's or log's own service first. See The Around this span and Around this log tabs.

See the whole grid at this time¶

At the top of the Around this span and Around this log tabs, select See the whole grid at this time. It closes the details panel and opens the Snapshot page with the time range picker set to the 15 minutes ending at the span's or log's time.

If the span or log line has no usable time, the link is not shown.

What the Snapshot page shows¶

From top to bottom:

1. The Snapshot heading, and the line: "How the grid was doing over the range selected above, and which services were running at its end."

2. The four charts, for the whole selected range:

   Chart	Shows
   Apdex (T=500)	Apdex over the range
   Request Rate	Overall request rate
   Error Rate	Overall error rate
   p90	p90 (90th percentile) over the range

3. What was running, taken at the end of the range.

What was running¶

A sentence above the list states the period it covers. Usually it reads:

Services that sent telemetry in the last 5 minutes of the selected range.

Below it, the list names each service that sent traces in those 5 minutes, sorted by service name. The list has one column, Service.

Read the list exactly for what it says:

• A listed service sent traces with at least one span (one step of a trace) that started in those 5 minutes. The list goes by when each span started, not by when it reached the grid.
• A service that is not listed sent no traces the grid could attribute to it in those 5 minutes. The list does not say why, and it does not say whether the service was up or down.
• The list is built from traces only. A service that sent only logs in those 5 minutes is not listed.

If the list was measured and no service sent traces, you see:

No service sent telemetry at this time.

Reading the list in practice¶

To check an earlier period, choose a custom range that ends 5 minutes earlier and compare the two lists. Make it any length other than exactly 15, 30 or 45 minutes: the picker treats a range of exactly those lengths as Last 15, 30 or 45 minutes, which ends now. Each list covers the 5 minutes before the end of its range, so the two lists do not overlap.

How the snapshot follows the time range picker¶

The time range picker is the only time control on the Snapshot page. The page has no time control of its own.

• The snapshot time is the end of the selected range. The charts cover the whole range. What was running is taken at its end.
• Change the range, and the page follows. Pick a new range in the picker and the charts and the list update to match.
• A link that names a time sets the picker. See the whole grid at this time sets the picker to the 15 minutes ending at that time, so the picker always shows the range the page is using.

The Snapshot page does not show the time as a row of its own. Read it from the time range picker, or, when the list is taken at an earlier time, from the sentence above the list. The Around this span and Around this log tabs show the time themselves.

When the end of the range is too recent¶

Data for a time keeps arriving for a few minutes afterwards. In this build, a time is treated as complete once it is at least 5 minutes in the past. This is an assumed delay, not a measurement of when the data has actually arrived, so very late data can still change the list after that. Before a time is treated as complete, the list for it could still change.

When you have just chosen a range such as Last 15 minutes, its end is the current time, so it is not treated as complete yet. Ranges such as Today or This Month end in the future. In these cases the page does not leave the list empty. It takes the list at the newest complete time instead, which is about 5 minutes before you opened the page or changed the range, and the sentence above the list names that time:

Services that sent telemetry in the 5 minutes up to time.

The time range picker does not move, and the charts still cover the whole selected range.

The page works out the newest complete time from your computer's clock and asks for it once. If the server does not treat that time as complete either, which can happen when your computer's clock is ahead of the server's, the list is not shown and you see:

This time is too recent

Data for this time is still arriving, so what is shown here can still change. Do not use it for an alert or an SLO decision yet.

What you can do: choose a range that ends further in the past. If the alert keeps showing, check that your computer's clock is set correctly.

The Around this span and Around this log tabs¶

These tabs show the snapshot for one span or one log line, with its own service first. The tab loads only when you select it.

At the top of the tab you see:

• The heading Around this span or Around this log.
• The line "What service and the rest of the grid were doing when this span was recorded." (or "this log was recorded").
• Time: the time the snapshot describes, to the second, in your own time zone with the zone named, for example 13 Oct 2026, 14:05:00 GMT+2. Hover over it to see the same time in UTC.

Below that, a heading with the service's name and the sentence:

Services that sent telemetry in the 5 minutes before this span.

The list shows every service that sent traces in those 5 minutes. The span's or log's own service comes first, tagged this span or this log. The rest follow, sorted by service name.

If the span's or log's own service sent no traces in those 5 minutes, the tab says so above the list:

service sent no telemetry in the 5 minutes up to this span.

The tab has no time control. To see the whole grid at that time, with the charts, select See the whole grid at this time.

If the span or log line is so recent that its time is not treated as complete yet, the tab shows the newest complete time instead, and Time shows which time that is. In that case the sentences in the tab describe the time shown in Time, not the span or log line itself.

Sharing a snapshot¶

The page address holds the selected time range, so copying the address shares the snapshot. What the other person sees depends on how the range was chosen:

To share a fixed time, use See the whole grid at this time, or a custom range that is not exactly 15, 30 or 45 minutes long, then copy the address.

Even with a fixed range, the other person can see something different from you when:

• The end of the range was too recent when you looked. You saw the list at the newest complete time. If they open the link later, the end of the range may be treated as complete by then, and they see the list at the end of the range itself. Check the sentence above the list: it names the period the list covers.
• The time has passed the grid's retention. They see that the data has been removed. See When the list cannot be given.
• They do not have access to the grid. They see "This time could not be loaded". See Limits.

The Around this span and Around this log tabs have no address of their own. To share what a tab shows, select See the whole grid at this time and copy the address of the page it opens.

When the list cannot be given¶

Sometimes the snapshot cannot say what was running. It then shows this in place of the list:

We cannot say what was running yet: reason

A service missing from here means we have no information about it, not that it was down.

The reasons you can see in this build:

For a time that is too recent, the snapshot shows the This time is too recent alert instead of a reason. See When the end of the range is too recent.

Data has been removed¶

For a time past the grid's retention, an alert appears above the list:

Data has been removed

This time is past the 30-day data retention. Its data is gone and cannot be recovered.

The Snapshot page shows this message as fixed text. The period in it is not read from your grid. What decides whether a time counts as removed is your grid's trace retention setting, so trust that setting over the number in the message. If your grid has no trace retention set, the snapshot assumes a short history, so older times can show as removed much sooner than the number in the message suggests.

A time counts as removed once it is older than that retention, less 20 minutes. The 20 minutes are a 15-minute window that the removal check looks back over from the time, plus a 5-minute safety margin, so a snapshot stops showing data slightly before that data is deleted. The list itself still covers only the 5 minutes before the time.

Limits¶

• Time control. The time range picker is the only time control. See How the snapshot follows the time range picker.
• Presence period. What was running always covers 5 minutes and is built from traces only. See What was running.
• Busy periods. See When the list cannot be given.
• Recent times. See When the end of the range is too recent.
• Retention. See Data has been removed.
• Access. You need access to the grid. Without it, the snapshot does not load and you see "This time could not be loaded".
• Request rate. Snapshot requests are rate limited per caller, under the same limit as trace and log queries. If you reach the limit, you see "This time could not be loaded", and an error notification may also appear at the bottom of the screen. Wait, then select Try again.

Messages you may see¶

Related¶