July did not deliver one tidy market story. It produced two speeds: activity in several developer-facing packages rose sharply over the four-week comparison, while the largest container images continued to add enormous absolute volume at their usual, much steadier pace. Both are useful signals, but they answer different questions.
This report compares Monday snapshots from 2026-06-29 and 2026-07-27. It looks across public package, community, and container signals. The aim is not to crown a winner; it is to separate a meaningful change in current package usage from the background accumulation of established infrastructure.
Executive Summary
The clearest July story is stronger developer-package activity without enough independent evidence to turn it into a broader adoption claim.
- Trino's npm package
trino-client, Azure Data Lake Storage's npm package@azure/storage-file-datalake, GoodData's PyPI packagegooddata-sdk, and OpenMetadata's PyPI packageopenmetadata-ingestionrecorded the strongest eligible package-download increases. These flow signals show higher weekly package activity than four weeks earlier, not whole-platform adoption. - PostgreSQL, Redis, and Memcached led absolute Docker Hub pull additions. Their modest percentage changes are better read as continuing operational scale than as a fresh breakout.
- Elasticsearch drew more reader attention in July. That interest did not coincide with a validated public-adoption mover, making it a focused investigation signal rather than evidence of a broader usage shift.
Public Adoption Signals: A Burst of Package Activity
The fastest-changing public signals came from packages that sit close to a developer's working environment. The npm package trino-client attributed to Trino and @azure/storage-file-datalake attributed to Azure Data Lake Storage more than doubled from their respective comparison points; the PyPI packages gooddata-sdk and openmetadata-ingestion also cleared the report's baseline filter by a wide margin.
| Tool | Package | Signal | Current value | 4-week change |
|---|---|---|---|---|
| Trino | trino-client | npm weekly downloads | 193,037 | +131,938 (+215.94%) |
| Azure Data Lake Storage | @azure/storage-file-datalake | npm weekly downloads | 165,342 | +108,179 (+189.25%) |
| GoodData | gooddata-sdk | PyPI weekly downloads | 45,686 | +27,453 (+150.57%) |
| OpenMetadata | openmetadata-ingestion | PyPI weekly downloads | 135,440 | +53,973 (+66.25%) |
The common thread is access: these packages are used to connect to, manage, or operate data systems. That makes the lift worth following, particularly for the Trino and Azure Data Lake Storage packages. Still, one four-week increase can reflect a release cycle, a dependency update, or a change in downstream usage. The evidence supports a change in package activity; it does not yet establish a lasting shift in how teams are choosing entire platforms.
Established Infrastructure Kept Adding Volume
Cumulative Docker Hub pulls tell a deliberately less dramatic story. Mature images added tens or hundreds of millions of pulls, but their percentage gains stayed small because their historical totals are so large. This is scale with momentum in the absolute sense, not evidence that the underlying technologies suddenly accelerated.
| Tool | Docker Hub image | Current pulls | 4-week change |
|---|---|---|---|
| PostgreSQL | library/postgres | 11,068,636,780 | +182,378,925 (+1.68%) |
| Redis | library/redis | 11,022,773,061 | +138,236,449 (+1.27%) |
| Memcached | library/memcached | 13,266,724,823 | +59,189,711 (+0.45%) |
| MySQL | library/mysql | 5,051,402,648 | +36,155,550 (+0.72%) |
This contrast is useful. July's package gains point to activity around specific developer interfaces, while the container data shows the familiar base layer of data infrastructure continuing to run at extraordinary scale. Neither metric is a proxy for market share, and they should not be compared as though they were.
Reader Interest: Elasticsearch Drew Attention
Elasticsearch drew more reader attention in July than in the previous comparison period. It was not among the report's validated public-adoption movers, so the useful reading is narrow: more practitioners were investigating Elasticsearch, not that its overall usage changed.
That divergence is worth tracking. Reader attention can surface a research cycle before public activity changes, but one month's attention alone is not enough to establish why it increased or whether it will persist.
What To Watch Next
- If
trino-clientand@azure/storage-file-datalakeremain elevated in the next comparison period, the case for sustained package activity will strengthen. A quick return to their prior levels would point toward a transient release or dependency effect. - If the large Docker Hub images keep adding similar absolute volume while percentage changes remain modest, that will reinforce the interpretation of steady infrastructure scale rather than a new acceleration.
- If Elasticsearch continues to draw reader attention and a public activity or product-evaluation signal also strengthens, the case for a broader shift will become stronger. If the attention fades without corroboration, it will look more like a short-lived research cycle.
Methodology
- The report uses the Monday snapshots from
2026-06-29and2026-07-27, a four-week comparison period. - Public package downloads are flow metrics; Docker Hub pulls are cumulative metrics. Minimum baselines prevent very small starting values from dominating percentage rankings.
- The evidence pipeline excludes questionable source mappings and repeated metric histories from publishable findings, while retaining them in the evidence audit.
- These are directional proxies for public activity and reader interest, not direct measurements of market share or technology selection.