Endpoint: GET https://margincalculator.angelbroking.com/OpenAPI_File/files/OpenAPIScripMaster.json (~37MB, Content-Length: 36952921 confirmed via HEAD)

Production symptom (Python, urllib.request.urlopen().read()): three consecutive attempts between 09:15 and 09:44 IST today all failed with http.client.IncompleteRead — the server closed the connection before delivering the promised Content-Length, with successively less data delivered each time:

• 09:15:04 — 648,242 bytes read, 36,304,679 more expected
• 09:36:31 — 572,061 bytes read, 36,380,860 more expected
• 09:43:37 — 393,216 bytes read, 36,559,705 more expected

Independently reproduced from a completely different network (not the machine the above ran on) using plain curl, tested over both IPv6 and IPv4-forced, same URL:

• 60s attempt: 4.25MB delivered, ~71KB/s
• 60s attempt (IPv4 forced): 2.93MB delivered, ~49KB/s
• 180s attempt: connection closed by the server at 3.5MB (~48KB/s) — well short of 37MB even at that slow rate
• Retested ~90 minutes later: 336KB in 60s, ~5.5KB/s — getting worse, not better
• Baseline sanity check on the same network in the same window: an unrelated 10MB file from a different host downloaded at 3.1MB/s, ruling out a local bandwidth problem on our end

So this isn't one flaky client or one flaky network — two different tools (Python urllib, curl), two different networks, and both IP protocols all show the same severe throttling against this one host, persisting for over an hour and trending worse.

This is a hard blocker — the scrip master is a prerequisite before any strategy can even start routing, so trading has been unable to start today because of this alone.

Anyone else seeing this on margincalculator.angelbroking.com today? Given the pattern of the last several days on the rate-limit thread above, I'd appreciate someone actually looking at server-side telemetry/CDN health for this endpoint rather than the usual silence.