@kapa.ai elementary_report.html it generated is 70mb, when serving with nginx it lags quiet a lot, any advice?

A 70 MB report is large but manageable. Here are the main levers to reduce its size and improve serving performance: 1. Reduce the data window (--days-back) The most effective single change. Default is 7 days — try 3 or even 1:

edr report --days-back 3

[Report size workaround] 2. Lower --executions-limit Controls how many historical test invocations are loaded per test. Default is 720 — reducing it cuts both memory and file size significantly:

edr report --executions-limit 100

[Memory tuning] 3. Disable passed test metrics Removes detail for anomaly detection tests that passed:

edr report --disable-passed-test-metrics true

4. Disable failed test samples Removes row-level samples shown for failed tests:

edr report --disable-samples true

5. Combine all options for maximum reduction

edr report \
  --days-back 3 \
  --executions-limit 50 \
  --disable-passed-test-metrics true \
  --disable-samples true

[Combined approach] 6. Enable gzip compression in nginx Since the report is a large static HTML file with embedded JSON, enabling gzip in nginx can dramatically reduce transfer size. The Elementary team has noted that gzip compression significantly reduces load times for large reports. [Large env context] This is a standard nginx configuration — add

gzip on;

and

gzip_types text/html application/json;

to your nginx config. 7. Last resort: full-refresh Elementary tables If the Elementary tables themselves have grown very large, a full-refresh reduces history (at the cost of losing old data):

dbt run --select elementary --full-refresh

[Full-refresh workaround] > Note: Reports tend to grow especially large with hourly anomaly tests, long training periods, or many custom dbt tests. If the static report approach remains problematic long-term, Elementary's cloud solution handles large environments better than static reports. [Large env context] 👍 0 button 👎 0 button ℹ️ Tip: Mention @elementary.ai in the thread for follow-ups or DM the app for 1-to-1 help.