Skip to main content
Question

Athena Extraction Timeout

  • September 14, 2026
  • 2 replies
  • 18 views
michael.hamme13
Level 1
Forum|alt.badge.img+6

Large history table extraction keeps hitting an Athena query timeout.

I'm stuck on a recurring issue when pulling a large history table through Data Integration → Extractions. No matter whether I build it in the visual editor or the SQL editor, the run dies after a long-running query with:

[Simba]AthenaJDBC An error has been thrown from the AWS Athena client. Query timeout

I already tried cutting the row count by joining the history table against a header table first, no change. The only version that actually finishes is one where I add a hard date cutoff, but that obviously leaves out a chunk of history I actually need.

Anyone dealt with this on larger Athena-backed extractions? Looking for a way around the timeout that doesn't mean sacrificing part of the dataset.

    2 replies

    Danieldo
    Level 6
    Forum|alt.badge.img+2
    • Danieldo
    • Level 6
    • September 14, 2026

    Two things that might be worth checking. If your WHERE clause uses TRY_CAST or SUBSTRING on a string date column, that could be blocking predicate pushdown — Athena can't use any indexing on those functions and falls back to a full table scan, which would explain why the timeout hits at roughly the same point every time. The other angle is partitioning the extraction instead of pulling everything in one go, so the load gets split into smaller chunks that each finish well within the timeout window.

    yingtao.gu13
    Level 6
    Forum|alt.badge.img+11
    • yingtao.gu13
    • Level 6
    • September 14, 2026

    Good points above. One more thing I'd add from experience: if the JOINs aren't strictly necessary during extraction, it might be cleaner to just pull the raw table into the Data Pool first and handle filtering and joins in downstream transformations instead. Doing too much inline during extraction can create unnecessary overhead, and separating the steps tends to make the whole pipeline easier to debug anyway. Might not solve the timeout on its own but worth considering as part of the overall setup.

    Help companies to achieve digital process excellence
    Community Code of ConductAccessibility statement