Kobo does not seem to log detailed reading session events when using the Kobalt reader

[awesomegithubusername]

Description

Kobo does not appear to log the following events (among others) in the ExtraData column (blob) of the event table when a book is opened and read using Kobalt:

• EventType 1020 — Session start (book opened)
• EventType 1021 — Session end (book closed)

These events are logged as expected when using the stock Nickel reader to open and read the same book.

Why this matters

I'm using these events to generate an HTML report of detailed reading activity, including:

• Reading sessions per book, with start and end times.
• Reading activity grouped by year and month.
• Daily reading sessions and usage summaries.
• Other statistics derived from individual reading sessions.

Without these events, reading sessions initiated through Kobalt cannot be reliably included in the report, resulting in incomplete reading statistics.

Steps to reproduce

1. Open a book using the stock Nickel reader.
2. Inspect the Kobo database's event table and verify that EventType 1020 and EventType 1021 are recorded in the ExtraData column.
3. Open and read a book using Kobalt.
4. Inspect the same table and check for the corresponding events.
5. Observe that the events are present for Nickel sessions but missing for Kobalt sessions.

Question

Is it possible for Kobalt to record these session start and end events in the Kobo database, or is there an alternative way to track reading sessions accurately when using Kobalt?

Thanks for your work on Kobalt!

[nicoverbruggen]

If it's an omission, I'd rather add it, yes! Absolutely.

[awesomegithubusername]

Thank you very much!

[nicoverbruggen]

This build should now write the correct data to the database, can you check whether that's the case?
https://kobalt-dev.nicoverbruggen.be/test-builds/2790d50fa4cd52348d75b29d972c8871a61eda389c4d3cb1

[awesomegithubusername]

I haven't checked the database yet, but a quick test confirmed that a last reading session done with Kobalt is indeed being calculated. Thanks a lot!

[nicoverbruggen]

Kobalt's database may now also contain useful info about reading sessions if Nickel doesn't give you sufficient info. Though that's not wired up in this build just yet, I should note. It only registers on Android at this time. The database table is there, though.

[peterboda236]

With the latest develop build the book closing time is really slow.

New build (Oct 29): book close takes ~3.9 s, almost all in the session step

From the logs:
{"event": "exit-requested", "utc": "2026-10-29T05:52:49.182Z"}
{"event": "native-save", "result": "verified", "utc": "2026-10-29T05:52:49.362Z"}
{"event": "reading-session-end", "events_saved": true, "milliseconds": 3112, "persisted": true, "seconds": 12, "time_saved": true, "utc": "2026-10-29T05:52:52.942Z"}
{"event": "profile-close", "elapsed_ms": 21, "utc": "2026-10-29T05:52:53.220Z"}
{"event": "book-close", "checkpoint_ms": 23.5, "native_ms": 43.4, "profile_ms": 25.1, "session_ms": 3133.5, "sync_ms": 3225.5, "total_ms": 3946.7, "utc": "2026-10-29T05:52:53.224Z"}
{"event": "session", "status": "closed", "utc": "2026-10-29T05:52:53.225Z"}

[nicoverbruggen]

Hmm, that's strange. I have a Clara BW too and it isn't that slow. Are you reading a very big book? Is your storage almost full? I'm curious what could cause this. Is it always slow?

[peterboda236]

Here's a short video opening and closing a few. One a bit slower than the others but I think all slower then it should be.

trim.9D18AFF2-8AD8-4D3B-BE86-E95486281161.MOV

Have 12 GB free space.

[peterboda236]

And this is beta 4

trim.BF44661E-3BFA-4CAB-88CE-D224B013A355.MOV

[peterboda236]

Ok, I've found it what is the issue here. If I use the built in sleep screen everything is fast in beta 5 too. If I use the Kobalt's sleep screen it's getting really slow.

[nicoverbruggen]

As I suspected... then it's likely the image being persisted to disk when the book is being closed that's slowing everything down. I'll see what I can do to make that faster.