In Ankit Verma’s account of debugging Ossian, a question about how retrieved documents and stored memories relate exposed two problems: citations counted chunks as separate sources, and memory recency kept aging from the original creation date. While building a Kafka Connect sink, he found a third: a delete failed on its first attempt because an ingest event referenced a document that had already been removed, but a retry took a different path and concealed the failure. The cases show why source identity, timestamp meaning, and first-attempt state changes deserve explicit tests.
Why chunk citations can overstate the number of sources
Retrieval often returns passages or chunks, but a reader may interpret each citation number as a separate source. In Verma’s Ossian setup, the system selected the top six chunks and numbered each one in the prompt. Three passages from engineering-handbook.txt appeared as citations [1], [2], and [3], alongside a passage from platform-architecture.md. That formatting could make one handbook look like three independent documents agreeing with one another.
Content-hash deduplication at ingestion did not solve this: the passages were distinct chunks in the same legitimate document, not duplicate files or identical content. The issue was the mismatch between the unit retrieved (a chunk) and the unit represented by a citation (a document).
Group evidence by document, not filename
Verma’s reported fix groups retrieved chunks by document ID before constructing the prompt. Each document gets one citation number, with its relevant passages kept together beneath it. Documents are ranked by their best-scoring chunk. Grouping by ID, rather than filename, also preserves the distinction between two different documents that happen to share a name.
Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstall#1 Best Overall
- Read Before You Buy — No Video Output: These adapters support charging and USB 2.0 data transfer, but cannot transmit video signals. Except for standard USB webcams (which use USB data only), they are not compatible with HDMI/DisplayPort cables, video-capable USB-C hubs, or docking stations with video output.
- Convert USB-A Ports to USB-C: Designed to connect USB-C earphones, cables, flash drives, card readers, and other USB-C accessories to standard USB-A ports. Plug-and-play with no drivers or software required.
- Aluminum Alloy Housing: Built with a sturdy aluminum alloy shell that aids in heat dissipation and protects against daily wear and scratches. Designed to maintain a stable and secure connection.
- Compact & Travel-Friendly: The ultra-compact design allows the adapter to stay plugged into your device without blocking adjacent ports or adding bulk, reducing wear and tear on your original USB ports.
- 12-Month Warranty: Backed by a 12-month manufacturer warranty for peace of mind. Designed to meet strict quality control standards for reliable everyday performance.
In one live question, six retrieved chunks became five citations, with two passages from one runbook under a single citation. That is one reported example, not a general measure of retrieval quality. The implementation details and outcome are Verma’s account, not an independently audited code review. Read Verma’s account on DEV Community.
Memory recency depends on what “recent” means
Ossian’s reported memory score used the formula similarity × importance × 0.5^(age / 30 days). The ranking query measured age from created_at. When a repeated statement matched an existing memory and an upsert updated updated_at, the ranking still treated the memory as if it had only been said once. A preference that was repeatedly confirmed therefore continued to decay from its original creation.
Rank #2
- 5-in-1 USB-C Hub: Experience comprehensive connectivity featuring a Power Delivery input, two USB-A 2.0 ports, a USB-A 3.0 port, and an HDMI port. (Note: The USB-C power delivery input port is only for connecting an external wall charger to power your laptop and cannot power peripheral devices.)
- 90W Pass-Through Charging: Achieve optimal charging with 90W pass-through power to your laptop, supported by a total input of 100W, with the hub reserving 10W for operational efficiency. (Note: Wall charger not included.)
- Quick Data Transfers: Accelerate your productivity with rapid data transfers using a high-speed 5Gbps USB 3.0 port and two 480Mbps USB 2.0 ports.
- 4K HDMI Display: Enhance your visual experience with a hub capable of delivering 4K resolution at 30Hz in both mirror and extend modes. Please note that this hub is compatible with MacBook (macOS 12 and newer), Windows 10 and 11, ChromeOS, and laptops equipped with DP Alt Mode and Power Delivery. Note: This device is not compatible with Linux.
- What You Get: Anker USB-C Hub (5-in-1, 4K HDMI), welcome guide, 18-month warranty, and our friendly customer service.
Separate creation, statement, and recall times
These timestamps answer different questions:
created_at: When was this memory first stored?- A last-stated or last-confirmed timestamp: When did the user most recently express this fact?
last_used_ator a similar recall timestamp: When was the memory retrieved or used?
If the goal is to rank facts by how recently they were stated, update the last-stated time when a deduplicating upsert confirms an existing memory, then use that time in the decay calculation. Simply refreshing a recall timestamp is not equivalent: Verma considered that approach and found that recalling both old and newer contradictory memories could refresh them together, leaving them tied on recency.
Verma reports that, in the running system, a fresh “switched the editor to the light theme” memory scored 0.765; “prefers the dark theme” scored 0.102 when 90 days old; and the dark-theme memory scored 0.817 after the preference was restated. These are project-specific examples reported in the September 13, 2026 article, not portable benchmarks.
Rank #3
- Sleek 7-in-1 USB-C Hub: Features an HDMI port, two USB-A 3.0 ports, and a USB-C data port, each providing 5Gbps transfer speeds. It also includes a USB-C PD input port for charging up to 100W and dual SD and TF card slots, all in a compact design.
- Flawless 4K@60Hz Video with HDMI: Delivers exceptional clarity and smoothness with its 4K@60Hz HDMI port, making it ideal for high-definition presentations and entertainment. (Note: Only the HDMI port supports video projection; the USB-C port is for data transfer only.)
- Double Up on Efficiency: The two USB-A 3.0 ports and a USB-C port support a fast 5Gbps data rate, significantly boosting your transfer speeds and improving productivity.
- Fast and Reliable 85W Charging: Offers high-capacity, speedy charging for laptops up to 85W, so you spend less time tethered to an outlet and more time being productive.
- What You Get: Anker USB-C Hub (7-in-1), welcome guide, 18-month warranty, and our friendly customer service.
Why a retry can hide a delete failure
While building a Kafka Connect sink for a Debezium-fed corpus, Verma reports that a delete removed a document and then attempted to write an ingest event containing that document’s ID. The event insert violated a foreign-key constraint because the referenced document no longer existed. The batch loop did not catch the error, so the batch failed.
The retry appeared to succeed, but it did not repeat the same state transition: the document was already gone, so the next delete found no document and the event used a null ID. The retry therefore concealed a repeatable first-attempt failure that happened once per document, according to Verma. The quoted backend error was: “insert or update on table "ingest_events" violates foreign key constraint Key (document_id)=(…) is not present in table "documents".” The ellipsis is a redaction in the published account.
Rank #4
- Dual Converters, Infinite Potential：Includes 2× USB C male to USB A female adapters and 2× USB A male to USB C female adapters. Perfect for a wide range of uses—tablets with Bluetooth keyboards, expand USB ports on macbook, and more. Two different converters for all your daily needs
- Next-Level 10Gbps & 3A Charging: No more slow 480Mbps, this usb to usb c adapter has a transfer speed of up to 10Gbps, allowing you to do more transferring in less time. This usb adapter fits both USB A and USB C charger, supporting up to 3A fast charging
- Upgraded Exquisite Craftsmanship: With an aluminum alloy housing and metal connector, the usbc to usb adapter is extremely durable and sturdy. Rigorously tested to withstand more than 10,000 times of plugging and unplugging, ensuring long-lasting performance
- Broad Compatible: The usb c to usb adapter widely supports all USB C/ USB A devices like laptops, tablets, cellphones, car chargers, and phone chargers. Such as compatible with MacBook Pro/Air 2023/2022, Thunderbolt 4/3 Devices,Apple MagSafe Watch 9/8/7/SE/Ultra, iPad Pro 2022/2021, Samsung Galaxy S23/S20/S10, and iPhone 17/16/15 Pro. Plug and play
- Please Note: To reach 10Gbps speed, keep the cable under 3.3 ft. For USB A Male to USB C adapters, try flipping the USB C connector. USB C Male to USB A adapters support bidirectional 10Gbps transfer within 3.3 ft
This pattern matters because retry success alone does not establish that the first attempt was safe. Inspect the state at each boundary: the document before deletion, the document after deletion, the event payload, and the foreign-key check. Tests should exercise the event API at the point where that constraint can fail, then verify both the initial failure path and the retry path.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Sink choices and the checks Verma reports
Verma describes the sink as using a stable, caller-supplied event ID because the event API is idempotent on that ID. The ID combines connector name, topic, partition, and offset with the record timestamp. The article says Debezium’s source position alone was unsuitable for initial snapshots because all rows in a snapshot share one LSN; using only that value could make distinct rows collide and later rows be discarded as duplicates.
Best Value
- 5-in-1 Connectivity: Equipped with a 4K HDMI port, a 5 Gbps USB-C data port, two 5 Gbps USB-A ports, and a USB C 100W PD-IN port. Note: The USB C 100W PD-IN port supports only charging and does not support data transfer devices such as headphones or speakers.
- Powerful Pass-Through Charging: Supports up to 85W pass-through charging so you can power up your laptop while you use the hub. Note: Pass-through charging requires a charger (not included). Note: To achieve full power for iPad, we recommend using a 45W wall charger.
- Transfer Files in Seconds: Move files to and from your laptop at speeds of up to 5 Gbps via the USB-C and USB-A data ports. Note: The USB C 5Gbps Data port does not support video output.
- HD Display: Connect to the HDMI port to stream or mirror content to an external monitor in resolutions of up to 4K@30Hz. Note: The USB-C ports do not support video output.
- What You Get: Anker 332 USB-C Hub (5-in-1), welcome guide, our worry-free 18-month warranty, and friendly customer service.
Delivery and record-handling behavior
- Blanked rows are removed so old text does not remain in the corpus.
- Records with none of the configured text fields are rejected as a likely configuration error, and Debezium placeholder values are refused.
- Synchronous
put()keeps offsets from advancing ahead of delivery. - Rate limits and 5xx responses back off using
Retry-After; rejected records go to a dead-letter queue; a 401 stops the task.
These are design details reported by the author, not independently verified findings. Verma also reports an end-to-end run against one Postgres table: a three-row snapshot became three answerable documents; changing a value from 180 to 90 days changed the answer without leaving an old chunk that still said 180; a delete removed the document and its chunks; a blanked row disappeared; and resetting sink offsets replayed 18 events before and after without adding documents. The “18 events before, 18 after” result is a project-specific replay check, not a benchmark.
Tests that challenge the implementation’s assumptions
“Every test passed. The tests had the same blind spots as the code,” Verma writes. The recency test backdated created_at, which was also the field the faulty query used. The test thus encoded the implementation’s mistaken assumption rather than distinguishing an old memory that had just been restated from one that had not.
Verma says a new test fails against the old query and that he checked this by restoring the old line. At the time of the sink bug, he reports that there were no tests for the event API. For systems with similar behavior, useful tests should distinguish last-stated from last-created and last-used timestamps, and should force a delete followed by an event write at the foreign-key boundary. A retry should be tested separately rather than treated as proof that the original operation was correct.
The Discord question that prompted the work was not fully resolved by these fixes. In Verma’s account, documents and memories remain separate; memories are not linked back to the documents from which they were learned, and semantically similar or contradictory memories are not reconciled. Grouping citations fixes how retrieved evidence is represented, and correcting recency fixes one ranking signal, but neither by itself reconciles conflicting stored memories.
Recommended Free Tools
Verma credits work with Claude (Anthropic) and says he checked the numbers and code samples against the running system. The article was reviewed and published by @dockndevai. Those disclosures are the author’s; the implementation and reported checks have not been independently reproduced here.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
