The options train is frozen. My algo set a limit sale of $3.5 on the 717 strike of QQQ. The sale has not executed and the dashboard has show the same bid and ask for 5 straight minutes while the last sale price has increased to 4.99.
Yes, the indicative feed is completely FUBAR. Freezes after every API purchase. Limits set for sale with API and are blown past for minutes to 10s of minutes. Unusable. I would not trade real money with the indicative feed today.
What is interesting is that I can watch the option chain on my dashboard and other bids and asks for other strikes are changing but not for the one I have an open position on and a limit sell order on.
I canceled my limit sell order and liquidated the position for the price that had been frozen for 15 minutes. Then within a minute the feed on the options chain dashboard started to move again.
The fact the chain data is all messed up and the https://status.alpaca.markets/ page shows everything is fine doesn’t instill a lot of confidence here.
It appears that 1DTE (next trading day so 3DTE) options are doing better, but I would not trust it today. No, the 3DTE options chains are also hanging up from time to time.
@DougFish @runin2k1 This morning, we experienced a brief but significant disruption in our OPRA options data feed shortly after market open, caused by heavy backpressure on the feed. During remediation, one of our streaming media components was restarted, which unfortunately caused downstream failures in our options quote streaming service. This led to some option quotes — including near-expiry (0DTE) contracts like the ones you mentioned — getting stuck and not updating for a period, resulting in the staleness you saw.
Our engineering team identified the root cause and has restored normal operation. We generally avoid restarts during market hours precisely because of this kind of risk, and this incident occurred while trying to resolve a separate, larger issue.
We’re continuing to monitor the feed and reviewing our incident response process to prevent a recurrence. Apologies for the inconvenience this caused, and please let us know if you noticed any other side effects (e.g., order issues) tied to this window so we can look into them specifically.
Thanks for the explanation Dan. I’m still seeing some degraded performance today with options quotes lagging quite a bit. Happy to send over additional details if it is helpful to your team.
Window: 2026-08-10 11:33:07 – 11:38:19 ET
Underlying: SPY @ ~773.57 · 0DTE expiry: 2026-08-10
Feeds: options feed=opra, equities feed=sip (Algo Trader Plus)
Contracts sampled: 8 near-ATM strikes
Summary
Near-ATM 0DTE SPY option quotes are arriving with their exchange timestamp 31s behind Alpaca’s own server clock (worst observed: 124s). SIP equity quotes for the same underlying, polled in the same loop over the same connection, are 0.4s — i.e. current. The problem appears specific to the options feed.
| Option quotes (OPRA) | Equity quotes (SIP) — control | |
|---|---|---|
| samples | 576 | 72 |
| median lag | +30.8s | -0.4s |
| worst lag | +124.0s | +0.7s |
| best lag | +10.0s | -1.0s |
2 Likes
im seeing the same today
It has been off since about 10am ET Friday 08/07. This feed seems to always have an open burst lag, but it hasn’t been recovering now.
More importantly – 30% of the trades feed isn’t reaching us at all — roughly 9 million trades missing in total since Friday 08/07
Throwing a +1 on this. Pulled some metrics off my platform in AWS US-East-1 (3ms from WS endpoint) and it lines up with the issues others are seeing. It looks like the paper trade endpoint may be running its fills based on the delayed feed as well.
Window: 2026-08-11 16:22:00–17:31:00Z (12:22–13:31 EDT) — the Alpaca-OPRA window
Quantity: receive time − exchange timestamp on the message
| Metric | SIP (Alpaca equities) | OPRA (Alpaca options) |
|---|---|---|
| samples | 3,793,126 | 55,610,150 |
| throughput | 916 /s | 13,432 /s |
| p50 (median) | 7.42 ms | 2.14 s |
| p95 | 10.48 ms | 24.9 s |
| p99 | 12.26 ms | 29.5 s |
| worst | no sample ≥ 30 s | > 30 s — unbounded |
| > 1 s | < 0.86% | 67.07% (37,295,790) |
| > 10 s | < 0.086% | 18.04% (10,032,076) |
| > 30 s | 0 (exact) | 0.54% (301,494) |
| mean | 8.63 ms | ≥ 4.78 s |
Dang, I see in Slack someone is posting the same thing and getting no response from the Alpaca team there either.
Hopefully the communication improves in the future, but this sort of issue with radio silence from the Alpaca side has been enough to scare me off the platform.
Would love to discuss this in Slack instead of the forum, but the slack link on the website doesn’t seem to work (Says ‘This link is no longer active’). Not sure where to report that for action. 
Still seeing the issue today, here is a fresh grab:
SIP vs OPRA — 2026-08-12 15:12:21–15:17:21Z (11:12–11:17 ET)
| Metric | SIP (Alpaca equities) | OPRA (Alpaca options) |
|---|---|---|
| samples | 386,607 | 2,796,961 |
| throughput | 1,289 /s | 9,323 /s |
| p50 (median) | 8.15 ms | 18.27 s |
| p95 | 11.02 ms | 33.62 s |
| p99 | 13.27 ms | 40.37 s |
| worst (max) | — | 44.39 s |
| best (min) | — | 4.36 ms |
I’ve been seeing the same issue since Friday (the 7th).
Interestingly the last three days (Mon, Tue, and Wed), the problem has solved itself after a couple hours of trading each day. Almost like there’s some overfilled queue in the mornings that needs to be cleared out, and then the feed works again.
@Dan_Whitnable_Alpaca This needs escalated please. Since August 7, Alpaca OPRA WebSocket, latest-quote REST, and snapshot REST data have repeatedly arrived 20–40 seconds behind their exchange timestamps, while simultaneous SPY SIP data remains current.
On August 12, one measured window showed:
OPRA WebSocket median: 31.6s; P95: 39.9s
Options REST latest median: 32.1s; P95: 42.7s
Options snapshots median: 32.8s
OPRA observations within two seconds: 0%
Simultaneous SPY SIP median: 276ms
+1 You guys need to stop wasting developer time with useless new features and focus on ensuring the API and streams work reliably. We spend so much time developing on this platform and these delays aren’t acceptable! You need to identify these and warn platform users that your service is degraded.
Looks SO MUCH healthier today! Are others seeing this too?
2026-08-13 — 13:35:00–13:40:00Z (09:35–09:40 ET)
| Metric | SIP (Alpaca equities) | OPRA (Alpaca options) |
|---|---|---|
| samples | 861,655 | 2,339,919 |
| throughput | 2,872 /s | 7,800 /s |
| mean | 10.41 ms | 34.94 ms |
| p50 (median) | 9.53 ms | 18.31 ms |
| p95 | 15.49 ms | 162.3 ms |
| p99 | 16.06 ms | 237.8 ms |
| worst (max) | — | ≤ 1.0 s |
| best (min) | — | 2.5 ms |
Big thanks to the team for tracking down whatever the issue was!!
1 Like
Yes looking back to normal today for me
1 Like
The OPRA option data latency issue has been resolved as of Wednesday, Aug 12. This was caused by an internal infrastructure incident affecting our OPRA data feed.
What happened: A deployment on Friday, Aug 7, introduced a performance issue in the OPRA feed consumer service, which caused the feed to get intermittently dropped and led to delayed or stale options quotes over the following few trading sessions.
Current status: The problematic deployment was rolled back on Wednesday morning, and the feed has been stable since — dropped messages returned to zero, and performance is back to normal. The incident has officially been marked as resolved, though the engineering team continues to monitor it closely.
Again, apologies for this, and thank you for your patience while it was being resolved!
2 Likes