Hi @Jagdish_j_ptl,
I don’t think there is any issue with live feed delay. I’m pasting below screenshot for info:
I took the same timestamp & instrument example of yours, here’s how to interpret:
uts => currentTs (a.k.a upstoxTimeStamp) coming as part of every payload
rts => receivedTimeStamp (system time, depends on which time server your OS is synced to, in my case since I’m using MacOS, it will be time.apple.com)
ltt => lastTradedTime (time traded at source)
Upstox themselves have 300-400 ms difference w.r.t LTT (which is on the higher side, for all computation & scaling/distribution reasons). If you look at rts & uts, rts says system received faster than what Upstox payload received, which cannot be (this proves that the time sync servers that we use have variations [it varies upto 500-600ms, generally on day-to-day, nobody cares about this])
In your case, there could be 2 possibilities:
- Your log prints are directly running on receiving thread and blocking the socket (which is getting queued up, probably) and causing more OS thread context switches.
- Your time sync server has couple of seconds variation compared to what NSE/Upstox server are using.
FYI, I’ve tested with streaming more than 3000 instruments and I don’t see any frustrating lag (only negligible like 50-60 ms, with CPU usage peaking hardly 3-5%). Check your CPU usage too while running your program.