Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Guides Qdrant data volume scaling decisions. Use when someone asks 'data doesn't fit on one node', 'too much data', 'need more storage', 'vertical or horizontal scaling', 'tenant scaling', 'time window rotation', or 'data growth exceeds capacity'.
.claude/skills/qdrant-qdrant-scaling-data-volume/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | -2% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -52% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -12% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | -50% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | -5% | 0% |
This document covers data volume scaling scenarios, where the total size of the dataset exceeds the capacity of a single node.
If the use case is multi-tenant, meaning that each user only has access to a subset of the data, and we never need to query across all the data, then we can use multi-tenancy patterns to scale.
The recommended way is to use multi-tenant workloads with payload partitioning, per-tenant indexes, and tiered multitenancy.
Learn more Tenant Scaling
Some use-cases are based on a sliding time window, where only the most recent data is relevant. For example an index for social media posts, where only the last 6 months of data require fast search.
Learn more Sliding Time Window
Most general use-cases require global search across all data. In these situations, we might need to fall back to vertical scaling, and then horizontal scaling when we reach the limits of vertical scaling.
When data doesn't fit in a single node, the first approach is to scale the node itself — more RAM, better disk, quantization, mmap. Exhaust vertical options before going horizontal, as horizontal scaling adds permanent operational complexity.
Learn more Vertical Scaling
When a single node can't hold the data even with quantization and mmap, distribute data across multiple nodes via sharding.
Learn more Horizontal Scaling
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 15,280 | 9,605 | -37% | 1 | 1 | 0% | 2,433 | 1,222 | -50% | 0 | 0 | — |
case-02 | pass→pass | 20,820 | 13,074 | -37% | 1 | 1 | 0% | 2,896 | 2,746 | -5% | 0 | 0 | — |
case-03 | pass→pass | 15,769 | 13,125 | -17% | 1 | 1 | 0% | 2,680 | 2,378 | -11% | 0 | 0 | — |
case-04 | pass→pass | 15,982 | 3,636 | -77% | 1 | 1 | 0% | 1,951 | 920 | -53% | 0 | 0 | — |
case-05 | pass→pass | 21,000 | 4,874 | -77% | 1 | 1 | 0% | 3,291 | 948 | -71% | 0 | 0 | — |
case-10 | pass→pass | 21,418 | 8,882 | -59% | 1 | 1 | 0% | 3,354 | 1,782 | -47% | 0 | 0 | — |
case-06 | pass→pass | 23,172 | 13,747 | -41% | 1 | 1 | 0% | 3,219 | 2,318 | -28% | 0 | 0 | — |
case-07 | fail→pass | 14,847 | 13,731 | -8% | 1 | 1 | 0% | 2,372 | 2,325 | -2% | 0 | 0 | — |
case-08 | pass→pass | 9,132 | 7,541 | -17% | 1 | 1 | 0% | 1,666 | 1,653 | -1% | 0 | 0 | — |
case-09 | pass→pass | 16,043 | 11,133 | -31% | 1 | 1 | 0% | 2,944 | 2,233 | -24% | 0 | 0 | — |
case-11 | fail→pass | 18,356 | 6,695 | -64% | 1 | 1 | 0% | 2,507 | 1,208 | -52% | 0 | 0 | — |
case-12 | fail→pass | 5,935 | 3,070 | -48% | 1 | 1 | 0% | 955 | 842 | -12% | 0 | 0 | — |
case-13 | pass→pass | 19,661 | 11,958 | -39% | 1 | 1 | 0% | 2,923 | 2,133 | -27% | 0 | 0 | — |
case-14 | pass→pass | 13,195 | 4,556 | -65% | 1 | 1 | 0% | 1,974 | 1,027 | -48% | 0 | 0 | — |
case-15 | fail→fail | 16,561 | 13,766 | -17% | 1 | 1 | 0% | 2,587 | 2,228 | -14% | 0 | 0 | — |
case-16 | pass→pass | 20,515 | 4,726 | -77% | 1 | 1 | 0% | 2,898 | 936 | -68% | 0 | 0 | — |
case-17 | pass→pass | 17,042 | 4,804 | -72% | 1 | 1 | 0% | 1,952 | 1,056 | -46% | 0 | 0 | — |
case-18 | pass→pass | 20,910 | 11,653 | -44% | 1 | 1 | 0% | 2,777 | 1,930 | -31% | 0 | 0 | — |
case-19 | pass→pass | 11,046 | 5,537 | -50% | 1 | 1 | 0% | 1,780 | 1,113 | -37% | 0 | 0 | — |
case-20 | pass→pass | 13,023 | 16,964 | +30% | 1 | 1 | 0% | 2,343 | 2,017 | -14% | 0 | 0 | — |
case-21 | pass→pass | 17,046 | 11,379 | -33% | 1 | 1 | 0% | 3,356 | 2,690 | -20% | 0 | 0 | — |
case-22 | pass→pass | 6,949 | 5,807 | -16% | 1 | 1 | 0% | 1,215 | 1,157 | -5% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +14 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.