Skip to content
dbexplore

Changelog

What we have published, and when

A dated log of the pages on this site: what went up, on which day, and what each one answers. It is a record of this site rather than of product releases.

What this page is, and what it is not

Everything below is something you can open and read. A calculator, a guide, a definition, an answer to a question people actually type. Each entry says what went up and why it was worth writing, and nothing is listed here that is not on the site.

It is not a release history for the product. DBExplore is in early access with design partners and has published no versioned releases, so there is nothing honest to list under that heading yet. Nothing below describes a product feature shipping. It describes a page being published, which is a different thing, and a changelog that let the two blur would be the first inaccurate sentence on a site that spends a lot of effort not having one.

Entries carry the day the work went up, and there are no month headings. The site went live on 16 September 2026, and rolling a fortnight of work up into months would tell you less than the dates already do.

Where it stands today

As this page is built, the site carries thirteen reference guides, fifty-four pages on what each PostgreSQL version changed under monitoring, forty glossary terms, sixty FAQ answers, six calculators, and six engineering posts. Those figures are counted from the pages at build time rather than typed, so they cannot drift from what is actually here.

Every SQL example in the guides, the glossary and the version pages was executed against a real PostgreSQL server of the major it names before it shipped, and the output committed beside the query rather than written from memory. That is the unusual thing about this site and the reason the pages below took as long as they did.

Published so far

Four guides deepened, and their SQL run on real servers

Three reference guides stopped being explainers and became references you could work from: autovacuum and table bloat, query plan regression, and pg_upgrade and extensions. A fourth is new: what an agent has to be able to see before anyone should let it act on a database.

In the same batch every SQL block under the guides was run against PostgreSQL servers in containers and its output committed next to it. That completed the set: 380 executed blocks across 105 pages, none of them typed out by hand.

The calculators got receipts

Every calculator now has its arithmetic replayed against real servers, one container per supported major, and the comparison committed: 453 predictions checked against what the server actually reported.

A calculator is a claim about somebody else's production database, which is a harder thing to be wrong about than a paragraph. The receipts are what make the claim checkable by someone who was not in the room when it was written.

Six calculators, sixty answers, and the landing pages

Free tools opened with six calculators and a check-up that runs all of them from one read-only query pasted back into the page: how long until writes stop, what an upgrade breaks in your monitoring, when autovacuum next visits a table, whether the settings fit in the RAM the server has, what each checkpoint costs, and how many connections to allow. Nothing you paste into any of them leaves the browser.

The same day, the FAQ opened with sixty questions in the words they get asked in, each answered in its opening paragraph; the glossary doubled to forty terms; the version pages reached fifty-four; and eight landing pages went up for the searches operators run when they are looking for a tool rather than an explanation, starting with the index advisor.

Reference, versions, and a glossary

Learn opened with twelve guides on the Postgres failures worth recognising on sight: what each one is, the query that finds it, and the change that ends it. The version pages opened covering majors 14 through 18, one page per change that moves a view, a column or a setting your monitoring reads. The glossary opened with twenty terms, and six more posts went up on the blog.

Put every Postgres you run on autopilot.

We onboard teams in small batches. Tell us about your fleet and we will reach out when a seat opens. One email, no drip campaign.

Request early access Book a demo