Ensure 100 % Trace Visibility Across All Time‑Range Buckets

Problem / Use-case

We learned that Instana rolls traces and spans into “time-buckets.” on serverside , for short and long term based traces, for operational technical reasons.

• Short presets such as Last 5 min or Last 15 min use very fine buckets (1-5 min).
• Longer presets such as Last 12 h, Last 24 h, or Last 7 d aggregate those same spans into much larger buckets

While that roll-up happens, the newest spans have not yet been copied into the larger buckets. 
As a result, engineers experience the following:

1. They filter in Trace Analyze by a Trace ID (e.g. call.traceId=abc123) and see it in Last 15 min.
2. They widen the time picker to Last 12 h—and the very same trace disappears for up to several hours.
3. Service and other Dashboards  show different call counts compared to Unbounded Analytics depending on the preset, so the data appears unreliable.

In day-to-day operations that behaviour breaks two essential workflows:

• Post-incident root-cause analysis – Analysts widen the window to 24 h to learn when an error began but conclude the trace “never happened” because it vanishes.
• Release comparison – SREs compare traffic patterns between today and yesterday, yet larger intervals under-report calls, hiding regressions.

Because Instana promises “every call, every service, always available,” the absence feels like data loss, not just delayed aggregation for the users, which weakens trust in the system.

Desired outcome

• A span that is visible a few minutes after ingestion in any short preset must remain visible in all longer presetslike 12 or 24h. 
• The Usage of short / long lived buckets is a technical solution, which should not be visible for the user. Dont try to solve it by  the UI which displays e.g. a brief indicator such as “ⓘ first/last bucket incomplete” instead of silently omitting data.

Business value

1. Lower MTTR – Engineers no longer waste 10–15 minutes re-querying smaller windows just to confirm data gaps.
2. Trustworthy dashboards – Trace Analze, SLO and KPI reports reflect the full set of calls, so teams stop exporting raw traces to external tools.

Acceptance criteria

1. A Trace ID that appears in Last 5 min (T + 2 min after ingestion) also appears in Last 12 h, Last 24 h, and Last 7 d within two minutes.
2. Total calls per service do not differ when switching between views for trace analyze traces view and services metrics.
3. The technical implementation is encapsulated in buckets is not visible and noticeable in the data for the User, Its a technical implementation which should not be visible or affect the users usage of the tool.
4. The enhancement is available in both Instana SaaS and Self-Hosted deployments.