Live source data, cached snapshots and missing values

Three different clocks affect a page

The upstream analytics service measures or calculates a dataset. This website then retrieves a copy. Finally, your browser loads the page. Those events can happen at different times, and refreshing the browser does not force the source to produce a new measurement period.

Read the dataset label and the retrieval status together. A snapshot obtained recently can still contain the source's most recent completed daily total. A daily aggregate should not be interpreted as a tick-by-tick feed of trades happening while you watch the screen.

Why the server caches data

A shared cache lets visitors read the same normalized snapshot without making a separate upstream request for every page view. It also reduces repeated downloads of large datasets. The server keeps data retrieval separate from private user actions such as saving a watchlist.

Caching introduces a trade-off: the local copy can lag behind the source. Its retrieval timestamp helps describe that lag. The timestamp is evidence that the server fetched a response, not proof that every underlying record was recalculated at that moment.

A stale value is different from a missing value

If a source refresh fails after an earlier successful retrieval, the previous snapshot can remain useful when labelled stale. The old retrieval time must remain visible. Replacing it with the time of a failed attempt would make the data appear fresher than it is.

If there has never been a valid response, the honest result is unavailable data. The site should not generate plausible numbers to fill the cards. Similarly, a missing field is not automatically zero. Zero is a numerical observation; unavailable means the observation cannot be supplied by the current dataset.

Check the comparison before drawing a conclusion

Two resources may refresh independently. If one table is current and another is stale, a comparison between them needs that qualification. A sudden difference might reflect the update gap rather than a change in activity.

An upstream correction can also change a historical value. Source adapters, classifications and price inputs can evolve. When reproducibility matters, record the source, measure, period and retrieval status alongside the value. A screenshot without that context may preserve appearance but lose the meaning of the observation.

What the source label means here

This installation retrieves analytics on the server and displays provenance information. It does not represent itself as an official DefiLlama service, and it does not convert analytics snapshots into executable exchange quotes. Paid API credentials, if configured later, belong on the server rather than in a browser request.

Use the data source page to find upstream documentation and the methodology to understand the site measurements. If a metric is unavailable, report that absence clearly. Accurate status is part of useful data, even when it produces an empty card instead of a dramatic number.