Issues tracker

If a team wants to make a batch download request through the portal that would be covered (free) on their subscription plan, they should be able to submit it regardless of the self-imposed team or user budget limits. Today, the user is unable to submit a request due to the "You've exceeded your team monthly limit. Your access to data is currently blocked." error. The workaround is to raise the team budget to at least $0.01 over the current monthly usage, as reported by https://app0.databento.com/portal/billing Internal tracking: D-3895

During winter months, when Eastern Time is UTC-4, the last hour of Nasdaq extended trading session (7 to 8pm ET) extends into the following UTC day. This is assigned D-2127 internally.

Currently our historical data discards Saturday test data on CME, however our live data gateways expose the test data because it’s useful for Databento’s own internal monitoring to ensure that our live gateways are running properly. We plan to discard this test data in live as well to ensure point-in-time behavior. We’ll likely have to hide test data from customers. This is assigned D-2036 internally and will likely only be addressed in Q3.

Currently, the InstrumentDefMsg provided by our definition schema does not contain the underlying leg information. Adding this is on our roadmap here. This is assigned internal ticket D-2032.

We currently leave out MDOrderPriority from our normalized MBO messages, due to 2 reasons: This behavior is CME-specific. Most other venues adopt ITCH-based behavior, where ordering is based on timestamp instead. If we supported MDOrderPriority, it would bloat the data for all of the non-CME venues. ts_event serves the same purpose as MDOrderPriority for all instruments except for interest rate options and instruments where there's a LMM. We confirmed this with CME GCC and also from practical experience of comparing simulation against live order matching. We actually recommend most users to infer MDOrderPriority from ts_event since it makes their order book implementation more reusable for other venues. Two proposed solutions were considered: Exposing raw message payloads and PCAPs. 'Supplementing' our normalized MBO data with venue-specific fields. We consider approach 2 to be inferior: If we tried incorporating other fields, we'd be replicating the venue's original structure in DBN with unnecessary indirection and copies and losing the benefit of normalization — so we might as well expose the raw message payload. For many venues, the raw data has nested message tree structures, and if we replicated it, either in Databento Binary Encoding or an unstructured encoding like JSON, we'd lose a lot of the benefits of our zero-copy binary encoding. The only limitation of approach 1 is that it's bandwidth-intensive, obviously as most venues require you to receive their raw multicast data on their extranet over a physical cross-connect. At this time, we reject approach 2 and are working towards approach 1. This is tracked on our roadmap here. This is tracked internally as D-2130.

Our CME MBO data assumes that, on a trade with a defined aggressor, there is only one order participating on the aggressor side (so if the Trade has side=A, at most one participating order will have a fill with side=A). This assumption does not hold in some cases. We have observed this behavior primarily on interest rate future spread trades. Quoting CME's documentation (https://cmegroupclientsite.atlassian.net/wiki/spaces/EPICSANDBOX/pages/46436881/MDP+3.0+-+Trade+Summary+Order+Level+Detail): "The aggressor quantity (tag 32) in the first Order Detail entry is not equal to the Summary Level fill quantity (tag 271) In this case, the aggressor joined a pool of resting orders and thereby created sufficient quantity to trigger a trade. This scenario can occur in any ratio spread where a different minimum quantity is required for each leg in order for the spread to trade. Examples include IVR and butterfly spreads. " This incorrect assumption leads to some Fill messages being reported with an incorrect side on the MBO schema. Here is an example of a such message: ts_recv,ts_event,rtype,publisher_id,instrument_id,action,side,price,size,channel_id,order_id,flags,ts_in_delta,sequence,symbol 2024-11-11T00:08:23.246556244Z,2024-11-11T00:08:23.243157109Z,160,1,42000697,T,A,110.093750000,4,19,8415051921726,0,22594,341221,ZNZ4 2024-11-11T00:08:23.246564192Z,2024-11-11T00:08:23.243157109Z,160,1,42000697,A,A,110.093750000,1,19,8415051921726,130,13621,341222,ZNZ4 2024-11-11T00:08:23.246837632Z,2024-11-11T00:08:23.243161063Z,160,1,42000697,F,B,110.093750000,1,19,8415051921726,0,17900,341225,ZNZ4 2024-11-11T00:08:23.246837632Z,2024-11-11T00:08:23.243161063Z,160,1,42000697,C,A,110.093750000,1,19,8415051921726,0,17900,341225,ZNZ4

This is a specific case of this other ticket: https://issues.databento.com/roadmap/incorrect-display-factor-used-for-the-strike-price-field-in-the-definition-schema-for-some-cme-options For now, use these correction factors to scale the strike price as a workaround (e.g. ECBTC's strike price needs to be multiplied by 100): EC6E 0.01 ECBTC 100.0 ECGC 10.0 ECHG 0.01 ECNG 0.1 ECO 10.0 ECSI 0.1 ECYM 100.0

e.g. 198,000.00 order here: 2024-09-15T12:00:05.054156821Z 2024-09-15T12:00:04.707396583Z 160 1 4358 A A 21643.000000000 1 8 6855758107682 40 0 2751 2024-09-15T12:00:05.054156821Z 2024-09-15T12:00:04.707396583Z 160 1 4358 A A 22000.000000000 3 8 6853082252068 40 0 2745 2024-09-15T12:00:05.054156821Z 2024-09-15T12:00:04.707396583Z 160 1 4358 A A 23500.000000000 1 8 6855315723493 40 0 2748 2024-09-15T12:00:05.054156821Z 2024-09-15T12:00:04.707396583Z 160 1 4358 A A 198000.000000000 1 8 6855095244654 168 0 2747

There are a few latency hotspots that we're working on. In general, the median latency is acceptable but we'd like to confine the tails. BBO-1x and OHLCV-1x There have been reports of latency spikes around our subsampled schemas, especially BBO-1x which would be used for a more latency-sensitive use case than OHLCV-1x and cannot be replicated from trades, sometimes stretching to a few seconds. These may be due to the memory bandwidth burst when publishing 600k to 1.2M instruments on the dot at the same time. Specific times of day Shared by one of our users, it appears that ts_out - ts_recv spikes to 1-3 seconds on certain days, e.g. attached screenshot for Sep 18, during Fed rate cut announcement. Mass cancels After stripping out a live gateway, we find that we're keeping to 5.1~6.5 us median, but 300-800 us 99.9th percentile. This appears to be due to latency spikes around mass cancels. dataset=GlbxMdp3 quantiles=[(0.0, 1565), (0.1, 2787), (0.25, 3597), (0.5, 5739), (0.9, 87679), (0.95, 214527), (0.99, 428031), (0.995..., (0.999, 658943)] dataset=GlbxMdp3 quantiles=[(0.0, 1615), (0.1, 2447), (0.25, 3187), (0.5, 5143), (0.9, 30079), (0.95, 68479), (0.99, 173567), (0.995, ..., (0.999, 312831)]

Since our site was originally designed with usage-based billing in mind only, there are parts of it like the Data usage page, pricing, and pricing calculator pages that might not correctly show $0 consumption fees once you're on a flat-rate plan like our Enterprise (soon renamed Unlimited) and upcoming Standard plans. If you discover an issue around this or have any confusion with pricing once you're on an flat-rate plan, do contact support for assistance. We're in the middle of auditing our site and making updates to improve UX for flat-rate subscription users. Most recently, the Billing page has been overhauled for this.

Due to our switches flip-flopping between an incorrect grandmaster, we occasionally see PTP issues at DC3. This is usually biased negative by at least -37 seconds because this leads to the TAI/UTC offset being momentarily before flipping over. In particular: Root cause - Switch loses sync with parent likely due to Cumulus/performance issues.Switch loses sync with its parent and enters holdover mode. In holdover mode, switch sets the PTP_UTC_REASONABLE flag to false.PTP NIC stops using the provided UTC offset of 37 and instead uses the default of 0.Outbound NIC by default respects the UTC offset valid flag. We've filed the issue with the switch manufacturer, but it's unlikely to be fixed—instead we'll be switching out to a different switch model in DC3 within September 2024. This issue should not affect our infrastructure in NY4 which has already been moved to the new switches. Based on our recent logs, we expect to see this at least twice per month. 08/06/24 02:17:05.035055 | 29780 | INFO | OSS | 19-8ff4c89e | Adapter 0 timestamp clock got Out-of-Sync with reference, TimeSkew: -55499999970 ns 08/21/24 07:31:54.314801 | 29780 | INFO | OSS | 19-8ff4c89e | Adapter 0 timestamp clock got Out-of-Sync with reference, TimeSkew: -55500001097 ns

Some information on system priced leg trades is available here: https://www.ice.com/publicdocs/System_Priced_Legs_Enhancement.pdf In practice, those can be very far off the best bid and best ask. These trades should be counted only for the volume, but not for the price.

A continuous contract with a statistics-based roll rule such as GC.n.0 can be incorrectly ranked if the exchange is late to publish statistics.

Our DBEQ.BASIC data has some records with instrument_ids that do not map to a symbol (or an empty symbol field). These all correspond to test symbols ATEST, CTEST, and PTEST, as well as subcategories of these such as PTEST W and ATEST C. Only test symbols are affected. Internal tracking number: D-2527

As CME sends the data, instruments with fractional pricing may have a display_factor that ends in e.g. 12.0 instead of 12.5. This caveat is noted on their page: EXCEPTION: For products that tick in modified fourths (30-Day Fed Funds options and Rough Rice options), the decimal '.5' is denoted in the price display. The Display Factor will be sent as '01' but should be treated as '00'. For example, an actual price of '12.5' would be displayed as '12' even though the display factor value '01' normally dictates a display of '12.5'. Therefore, client systems must include the .5 in the fractional conversion to decimal calculation for these products. Databento should apply this rule to the display_factor field for customer's convenience. Internal tag: D-2693