该用户的聊天记录定位。
thanks for reporting! working on a refactor to support this approach now. out of curiosity, how many simultaneous connections are you running?
thanks for reporting! working on a refactor to support this approach now. out of curiosity, how many simultaneous connections are you running?
that's how i'm reading it, too, but i've been misled by imprecisely worded documentation before and am asking about the actual server behavior
that's how i'm reading it, too, but i've been misled by imprecisely worded documentation before and am asking about the actual server behavior
under these new limits, is an `add_markets` with 10000 `market_tickers` one command or 10k commands?
under these new limits, is an `add_markets` with 10000 `market_tickers` one command or 10k commands?
https://docs.kalshi.com/changelog#june-18-2026-3 "Max 500k market subscriptions per session" / "Max 10k/s commands per second enforced" guess i'm adding sharding this weekend
https://docs.kalshi.com/changelog#june-18-2026-3 "Max 500k market subscriptions per session" / "Max 10k/s commands per second enforced" guess i'm adding sharding this weekend
fee rounding reinbursement showed up on my account today, thanks for taking care of this!
fee rounding reinbursement showed up on my account today, thanks for taking care of this!
similar context here, except i only subscribe to MVE-related tickers on this connection (so 95%+ instead of 100%). explicitly tries to avoid accumulating recv buffer, worked great until yesterday. looking forward to hearing your experience with this approach
similar context here, except i only subscribe to MVE-related tickers on this connection (so 95%+ instead of 100%). explicitly tries to avoid accumulating recv buffer, worked great until yesterday. looking forward to hearing your experience with this approach
dug up some old logs with "Channel error" and it was code 10 instead of 25: `{"code":10,"msg":"Channel error"}`
dug up some old logs with "Channel error" and it was code 10 instead of 25: `{"code":10,"msg":"Channel error"}`
Kalshi
维护后 WebSocket 出现 Subscription buffer overflow 错误 (code 25)
backpressure can occur either on kalshi's end or on our end, and it can be hard to distinguish the two. i believe i have a pretty good handle on it on my end, and have taken serious effort to process incoming messages promptly and without blocking. maybe this new error message vs "Channel error" encodes some information about the source of the backpressure. or maybe it has nothing to do with backpressure at all. i am just seeking some official word about whether websocket client expectations have changed, and what those new expectations are, before speculatively sinking a bunch of effort into sharding my consumption across multiple connections
backpressure can occur either on kalshi's end or on our end, and it can be hard to distinguish the two. i believe i have a pretty good handle on it on my end, and have taken serious effort to process incoming messages promptly and without blocking. maybe this new error message vs "Channel error" encodes some information about the source of the backpressure. or maybe it has nothing to do with backpressure at all. i am just seeking some official word about whether websocket client expectations have changed, and what those new expectations are, before speculatively sinking a bunch of effort into sharding my consumption across multiple connections
right. prior to this release, i would get "Channel error" during backpressure instead of "Subscription buffer overflow" and it happened much less frequently. it seems that they've changed something
right. prior to this release, i would get "Channel error" during backpressure instead of "Subscription buffer overflow" and it happened much less frequently. it seems that they've changed something
Kalshi
维护后 WebSocket 出现 Subscription buffer overflow 错误 (code 25)
i wasn't receiving skipped sequences before, it seems likely they've changed a buffer size on their end with this release. i do subscribe to a large number of markets, requiring me to start sharding would not be unreasonable. it would be nice to have official guidance about what this error means specifically
i wasn't receiving skipped sequences before, it seems likely they've changed a buffer size on their end with this release. i do subscribe to a large number of markets, requiring me to start sharding would not be unreasonable. it would be nice to have official guidance about what this error means specifically
Kalshi
维护后 WebSocket 出现 Subscription buffer overflow 错误 (code 25)
getting a LOT of these on orderbook_delta today ``` 2026-06-11T02:27:32.871454-07:00 recv {"type":"error","sid":5,"seq":788175,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T02:37:48.037057-07:00 recv {"type":"error","sid":5,"seq":749511,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T06:27:46.255354-07:00 recv {"type":"error","sid":5,"seq":1159130,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T07:28:02.518391-07:00 recv {"type":"error","sid":5,"seq":977122,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T12:46:50.11565-07:00 recv {"type":"error","sid":5,"seq":2603712,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T12:52:09.495339-07:00 recv {"type":"error","sid":5,"seq":1135716,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T12:58:00.776629-07:00 recv {"type":"error","sid":5,"seq":1161842,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T13:05:25.052673-07:00 recv {"type":"error","sid":5,"seq":1175098,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T13:23:00.852754-07:00 recv {"type":"error","sid":5,"seq":1235265,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T14:11:02.185442-07:00 recv {"type":"error","sid":5,"seq":1492122,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T14:17:27.32136-07:00 recv {"type":"error","sid":5,"seq":1203785,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T15:12:12.074937-07:00 recv {"type":"error","sid":5,"seq":1504756,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T15:17:17.270374-07:00 recv {"type":"error","sid":5,"seq":1205601,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T15:24:10.790776-07:00 recv {"type":"error","sid":5,"seq":1246885,"msg":{"code":25,"msg":"Subscription buffer overflow"}} ```
getting a LOT of these on orderbook_delta today ``` 2026-06-11T02:27:32.871454-07:00 recv {"type":"error","sid":5,"seq":788175,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T02:37:48.037057-07:00 recv {"type":"error","sid":5,"seq":749511,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T06:27:46.255354-07:00 recv {"type":"error","sid":5,"seq":1159130,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T07:28:02.518391-07:00 recv {"type":"error","sid":5,"seq":977122,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T12:46:50.11565-07:00 recv {"type":"error","sid":5,"seq":2603712,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T12:52:09.495339-07:00 recv {"type":"error","sid":5,"seq":1135716,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T12:58:00.776629-07:00 recv {"type":"error","sid":5,"seq":1161842,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T13:05:25.052673-07:00 recv {"type":"error","sid":5,"seq":1175098,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T13:23:00.852754-07:00 recv {"type":"error","sid":5,"seq":1235265,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T14:11:02.185442-07:00 recv {"type":"error","sid":5,"seq":1492122,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T14:17:27.32136-07:00 recv {"type":"error","sid":5,"seq":1203785,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T15:12:12.074937-07:00 recv {"type":"error","sid":5,"seq":1504756,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T15:17:17.270374-07:00 recv {"type":"error","sid":5,"seq":1205601,"msg":{"code":25,"msg":"Subscription buffer overflow"}} 2026-06-11T15:24:10.790776-07:00 recv {"type":"error","sid":5,"seq":1246885,"msg":{"code":25,"msg":"Subscription buffer overflow"}} ```
submit more taker orders imo
submit more taker orders imo
65,229 as of my most recent GetMarkets walk
65,229 as of my most recent GetMarkets walk
(would happily press the big red button if i learned that this was causing problems, though!)
(would happily press the big red button if i learned that this was causing problems, though!)
there can be 2M+ open MVE markets at once these days - during busy times i've had order counts in the five digits and could push it further if i had a higher rate limit to be able to safely manage them with. and i'm not doing 1c stink bids. not at risk of exceeding the 200k limit but i'm glad i'm aware of it now
there can be 2M+ open MVE markets at once these days - during busy times i've had order counts in the five digits and could push it further if i had a higher rate limit to be able to safely manage them with. and i'm not doing 1c stink bids. not at risk of exceeding the 200k limit but i'm glad i'm aware of it now
https://docs.kalshi.com/api-reference/orders/create-order "Each user is limited to 200 000 open orders at a time"
https://docs.kalshi.com/api-reference/orders/create-order "Each user is limited to 200 000 open orders at a time"
believe it or not, that's v1
believe it or not, that's v1
this was initially empty for me too, but has started reporting actual data now
this was initially empty for me too, but has started reporting actual data now
i'm fully migrated to v2, so no sweat disabling it at this point, just noticed the change and was curious what happened. thanks for the info!
i'm fully migrated to v2, so no sweat disabling it at this point, just noticed the change and was curious what happened. thanks for the info!
https://external-api.kalshi.com/trade-api/v2/account/endpoint_costs did the create order v1 costs decrease? <:JPOW:1400965831701303406>
https://external-api.kalshi.com/trade-api/v2/account/endpoint_costs did the create order v1 costs decrease? <:JPOW:1400965831701303406>
Kalshi
维护后 WebSocket 出现 Subscription buffer overflow 错误 (code 25)
got a few `{"code":25,"msg":"Subscription buffer overflow"}` errors in the hour after maintenance ended, that's a new one to me
got a few `{"code":25,"msg":"Subscription buffer overflow"}` errors in the hour after maintenance ended, that's a new one to me
portfolio combo tab on web no longer links to individual KXMVE* market pages 😭
portfolio combo tab on web no longer links to individual KXMVE* market pages 😭
when i see arguments that latency doesn't matter, i assume the person making them is trying to defend their latency arb strategy from potential newcomers, not trying to provide truthful advice
when i see arguments that latency doesn't matter, i assume the person making them is trying to defend their latency arb strategy from potential newcomers, not trying to provide truthful advice
i found the fill notifications with inverted prices really alarming for about a day, then got used to it. it was annoying to process the change and you'll find lots of similar sentiment from others in scrollback, but it seems unlikely that they're gonna roll it back at this point
i found the fill notifications with inverted prices really alarming for about a day, then got used to it. it was annoying to process the change and you'll find lots of similar sentiment from others in scrollback, but it seems unlikely that they're gonna roll it back at this point
are may fee rounding rebates still happening? for april's, i received it the same day i received my interest credit. i received may's interest credit yesterday, but no sign of a fee rounding rebate this time
are may fee rounding rebates still happening? for april's, i received it the same day i received my interest credit. i received may's interest credit yesterday, but no sign of a fee rounding rebate this time
yes, i use it to discover MVE markets to trade on
yes, i use it to discover MVE markets to trade on
crazy number of MVE markets open right now bc of world cup, ~1.4M as of my most recent GetMarkets walk. guess this is the new normal for a couple weeks
crazy number of MVE markets open right now bc of world cup, ~1.4M as of my most recent GetMarkets walk. guess this is the new normal for a couple weeks
it would be interesting to have markets on the results of public company shareholder votes
it would be interesting to have markets on the results of public company shareholder votes
i think this is the latest word https://discord.com/channels/871819895443189862/927686720990892032/1511116938833559592
i think this is the latest word https://discord.com/channels/871819895443189862/927686720990892032/1511116938833559592
can't seem to use the new `is_block_trade` filter https://docs.kalshi.com/changelog#may-29-2026 - queries always timing out when it's set to true ``` % curl -s 'https://external-api.kalshi.com/trade-api/v2/markets/trades?is_block_trade=true' {"error":{"code":"context_deadline_exceeded","message":"context deadline exceeded","service":"query-exchange"}} ```
can't seem to use the new `is_block_trade` filter https://docs.kalshi.com/changelog#may-29-2026 - queries always timing out when it's set to true ``` % curl -s 'https://external-api.kalshi.com/trade-api/v2/markets/trades?is_block_trade=true' {"error":{"code":"context_deadline_exceeded","message":"context deadline exceeded","service":"query-exchange"}} ```
it's the weekly maintenance window, thursdays 3-5 AM ET. usually they just disable trading, but sometimes (like today) they take the whole exchange down. it should be back in 1.5hrs
it's the weekly maintenance window, thursdays 3-5 AM ET. usually they just disable trading, but sometimes (like today) they take the whole exchange down. it should be back in 1.5hrs
scam link
scam link
if the 17c had arrived before the 16c, the 16c would have failed, which would have cascaded into the subsequent strikes. i could reformulate this into a single take @ 23c instead of 4 separate ones at different strikes, but that could potentially result in a larger total cost than anticipated if the orderbook changed under me while the order was in flight
if the 17c had arrived before the 16c, the 16c would have failed, which would have cascaded into the subsequent strikes. i could reformulate this into a single take @ 23c instead of 4 separate ones at different strikes, but that could potentially result in a larger total cost than anticipated if the orderbook changed under me while the order was in flight
guess i'll fire them off in parallel, pray they arrive in the order i want, and deal with the mess 1 rtt later if they're rearranged in a way that impacts execution
guess i'll fire them off in parallel, pray they arrive in the order i want, and deal with the mess 1 rtt later if they're rearranged in a way that impacts execution
looking over my order management code, this is only important to me for placement, if batch cancel went away it would be a simple change
looking over my order management code, this is only important to me for placement, if batch cancel went away it would be a simple change
sometimes i have a preferred order i'd like them executed in (ascending price order taking multiple strikes on a single market, e.g.), and submitting them in one batch gives me control over that, but sending many requests in parallel could result in them arriving and being executed out of order
sometimes i have a preferred order i'd like them executed in (ascending price order taking multiple strikes on a single market, e.g.), and submitting them in one batch gives me control over that, but sending many requests in parallel could result in them arriving and being executed out of order
kalshi amended this filing, switching from a request for cftc certification to a self-certification, accidentally(?) revealing a previously-withheld "confidential" appendix, and adding that btc perps "will initially be listed after close-of-business on June 2, 2026" https://www.cftc.gov/filings/ptc/ptc0602265036.pdf
kalshi amended this filing, switching from a request for cftc certification to a self-certification, accidentally(?) revealing a previously-withheld "confidential" appendix, and adding that btc perps "will initially be listed after close-of-business on June 2, 2026" https://www.cftc.gov/filings/ptc/ptc0602265036.pdf
view source
view source
a while ago i played with scraping the first party cfb websocket and crunching the window average numbers locally. got it to match up with the numbers on the kalshi ui, never developed it into an automated strategy bc (1) technically would violate licensing and (2) it's a crowded market and i figured others were already doing it. this probably destroys any edge that would have had
a while ago i played with scraping the first party cfb websocket and crunching the window average numbers locally. got it to match up with the numbers on the kalshi ui, never developed it into an automated strategy bc (1) technically would violate licensing and (2) it's a crowded market and i figured others were already doing it. this probably destroys any edge that would have had
no idea if it actually works yet, gotta plumb the index_id stuff before i can test subscribe, and not urgently motivated because i don't currently trade those markets
no idea if it actually works yet, gotta plumb the index_id stuff before i can test subscribe, and not urgently motivated because i don't currently trade those markets
whoa, when was this added? it's not in the asyncapi yaml spec... https://docs.kalshi.com/websockets/cfbenchmarks-value
whoa, when was this added? it's not in the asyncapi yaml spec... https://docs.kalshi.com/websockets/cfbenchmarks-value
https://docs.kalshi.com/websockets/public-trades
https://docs.kalshi.com/websockets/public-trades
strc 96c, wonder what's the blast radius of a depeg
strc 96c, wonder what's the blast radius of a depeg
i reduced the 2x2 down to book side on ingest before, because i sometimes place sell orders via the ui, but switched over to just reading the book side directly when i migrated my order entry over to v2
i reduced the 2x2 down to book side on ingest before, because i sometimes place sell orders via the ui, but switched over to just reading the book side directly when i migrated my order entry over to v2
it's going to be absolute chaos when action/side are removed from GetOrders
it's going to be absolute chaos when action/side are removed from GetOrders
was the exchange sharding plan scuttled? this page 404s now https://docs.kalshi.com/getting_started/exchange_sharding
was the exchange sharding plan scuttled? this page 404s now https://docs.kalshi.com/getting_started/exchange_sharding
rfqs are a distraction unless you're discriminating on creator_id. working directly on orderbooks gives you fill priority over rfq-ers, and with some care can be done comfortably within normal rate limits
rfqs are a distraction unless you're discriminating on creator_id. working directly on orderbooks gives you fill priority over rfq-ers, and with some care can be done comfortably within normal rate limits
iirc unauthenticated requests are charged against an (undocumented, smaller) per-ip bucket
iirc unauthenticated requests are charged against an (undocumented, smaller) per-ip bucket
with us dollars, which the federal government can create an infinite number of out of thin air
with us dollars, which the federal government can create an infinite number of out of thin air
life insurance = death markets
life insurance = death markets