Exposure Checks only show 100 entries (checks of the last 7 days)

[TeRe20200]

Hi everyone, just by accident, I checked in my Exposure-Checks on the app and the last date logged is the 14th. I checked with my friends and my parents too, 80% of them have only logged in the last 7 days, 20% have 14 days listed. We are taking about mostly Apple devices here (but also there are two Androids that are not showing last 14 days).

What shall we do? How can we solve this issue?

We do have the latest App-Update (Version 1.1.1) and those having only logged the last 7 days either updated on iOS 13.6, some didn’t.

I’d be really thankful for some input and help.

---

Internal Tracking ID: EXPOSUREAPP-1935

[TeRe20200]

Just as a little additional information: My app just updated the exposure-check for today, now the oldest exposure-check in the list is the 15th (while yesterday it was the 14th). Can anyone help?

[SebastianWolf-SAP]

Let's first clarify if we are talking about the same thing as the explanation was a little unclear. You are talking about the overview of the exposure checks which can be accessed via Settings -> Personal Data -> Health -> COVID-19 Exposure Logging -> Exposure Checks and that the earliest (not the last) date which is displayed there is July 14th, right?

I just also checked my log there and it truncated as well before the mentioned 14 days. For me it's July 12 - and only 5 entries for this date. I'm pretty sure that I also had 14 checks at this date...

After I checked the export of this log (the button almost at the bottom of the list) - I found some strange coincidence: The log shows exactly 100 entries. As I checked my logs daily and I definitely had 14 checks on July 12 and also the days before, this might be a bug in the list that we need to address to Apple. They might limit the number of entries to 100.

Could you please check if you and your friends also have 100 entries on their iPhones?

Mit freundlichen Grüßen/Best regards, SW Corona Warn-App Open Source Team

[TeRe20200]

Hi Sebastian, thank you for your time and answer:

Yes, Settings -> Personal Data -> Health -> COVID-19 Exposure Logging -> Exposure Checks brings me to a list that starts with today’s date and ends with the 15th of July (yesterday it ended with the 14th of July).

And I just counted:

• there are exactly 100 entries! (again from the 15th of July, ending with todays date)

Does that mean it can only show 100 entries but is still working with the last 14 days information?

[neonhaxred]

Hello there. I found this thread via google. I am experiencing the very same problem on iOS 13.6. With today my oldest entry is 17.07.2020 – yesterday it was 14.07.2020.

I exported the logs and read the JSON-File and there are exactly 100 entries as you predicted. With today the oldest entry in the JSON file is 2020-07-14. There are no records for the 15th, it goes on with the 16th. This behavior is strange insofar as the phone is always on, Bluetooth is always on, I was in range of at least two people with the app every single day and have never flushed the cache nor did anything else (not even activating flight mode or so).

[SebastianWolf-SAP]

OK, also here we have a confirmation by several people. In this case it's confirmed. We will address this to Apple and keep the ticket open until it's resolved.

@TeRe20200 That's correct! This list only shows the check history. As long as you have 14 checks in the last 24 hours everything is working correctly!

Everybody else: You can verify that everything works correctly by counting the number of checks within the last 24 hours. This number should match (or be 1 less) the number of active days. So if you had the app running already in the last 2 weeks or longer, you should always see 14 checks (or sometimes 13 depending on some boundary conditions) in the last 24 hours.

[SebastianWolf-SAP]

And completely forgot: Thank you very much for reporting this, @TeRe20200! That could explain some strange feedback which we have received over the last week and that we couldn't categorize or find our properly! :)

[neonhaxred]

I do count 26 counts only for today. Just want to provide this information.

[TeRe20200]

And completely forgot: Thank you very much for reporting this, @TeRe20200! That could explain some strange feedback which we have received over the last week and that we couldn't categorize or find our properly! :)

You are very welcome. Thank you for your help and reply. I hope you can help @neonhaxred too, that seems to be a different problem there.

[SebastianWolf-SAP]

@neonhaxred 26 for one day seems to be a little too much... But probably we address that in a separate thread. Maybe there was a second check of the last 2 weeks (2 x 13 days) and thanks to iOS 13.6 we don't have that problematic rate limiting anymore. Did you update/reinstall the app today?

[neonhaxred]

I did not restart the tracing nor did I flush the data. I always updated iOS to its latest version + the app.

Thanks for your kind replies btw!

[Mihahn]

Can confirm the behaviour on iOS 13.6. For me it's July 25th until July 31st – exactly 7 days and 100 entries.

Am I right that – by limiting it to 100 entries – all exposure data before the 25th got lost and if I had a match before the 25th (as it should be 14 days of logging) I can't be notified? That said the right information is not in the cwa (14/14 days) but in Apple's Privacy -> Health -> Exposure logging?

[neonhaxred]

Am I right that – by limiting it to 100 entries – all exposure data before the 25th got lost

That's what no one knows for sure. It seems like the app is still working finde due to the fact that it seems like you just need to ensure that there are 14 checks in the last 24 hours (a number that you easily do exceed). This is not yet confirmed though

[tkowark]

Please keep in mind, that we are talking about the exposure check log. Limiting this to 100 entries only means that you cannot see earlier checks of the log. As noted in the screen before (Covid-19-Begegnungsaufzeichnungen), the collected RPIs are stored for 14 days. So, when the app does its daily check, it checks against the entire 14 days of stored RPIs.

The 14/14 (or now: Dauerhaft Aktiv) days are also independent of that, they just denote for how long exposure logging has been active on your device without being shut off by manual deactivation, putting the phone into flight mode, etc.

[Mihahn]

Please keep in mind, that we are talking about the exposure check log. Limiting this to 100 entries only means that you cannot see earlier checks of the log. As noted in the screen before (Covid-19-Begegnungsaufzeichnungen), the collected RPIs are stored for 14 days. So, when the app does its daily check, it checks against the entire 14 days of stored RPIs.

Thanks for clarifying!

[Tho-Mat]

@SebastianWolf-SAP Any news from Apple about the 100 entries limit and when it will be removed? The issue is now 25day old.

[thomasaugsten]

We forward this to Apple and Google and they still discussing if they can increase the value because of data protection and privacy reasons. We have no direct influence on the OS settings

[Tho-Mat]

@thomasaugsten Thanks for your quick answer. Maybe I'm to stupid, but i do not see any data protection and privacy reasons here.

If the 14 day limit comes into account all protocol entries, except the last 13/14, have to be removed, since all other contain contact checks for dates that are more than 14 days old.

[neonhaxred]

Apple did not understand the problem yet. I sent them a screenshot of the problem once again and they said they're investigating

[thomasaugsten]

The issue is here some other apps using per check only one or two files this means with 100 entries you can look back 100 days.

[daimpi]

Why don't they (Apple/Google) make sure that they delete all entries older than 14 days, but show the ones which are more recent than that no matter how many entries there are?

[Ein-Tim]

With iOS 13.7, this problem is solved because now you find "all files in one file" See the Screenshots:

[daimpi]

Nice 😊

How does the exported .json file look like, do you maybe wana share yours? 🙂

[Ein-Tim]

@daimpi For sure:

ExposureChecks-2020-09-02.zip

[daimpi]

@GPclips I think this issue can be closed since this problem seems to have been resolved by the release of iOS 13.7 / 14.

[PhilippBerl]

I have the same problem: CWA 1.3.2., iOS 13.7.

CWA showed me first time last Friday 03.10.2020 "1 Begegnung mit niedrigem Risiko", yesterday 05.10.2020 "2 Begegnungen mit niedrigem Risiko" I am living alone, spending 95% of my time alone in my flat. I had no personal contact in the last 14 days that would meet the conditions to be in any way risky. (In der Realität kein einziger Kontakt unter 2 Meter länger als 5 Minuten.)

Although CWA is reporting 2 contacts with low risk, the iOS protocol of the exposure logs of the last 14 day shows 0 matches.

Is this a problem of iOS or of CWA? Which system is the determining system? Background of the question: Is the report in CWA false positive or is the report in iOS false negative?

Best wishes from Berlin ExposureChecks-2020-10-05 Kopie.pdf

[neonhaxred]

I am living alone, spending 95% of my time alone in my flat. I had no personal contact in the last 14 days that would meet the conditions to be in any way risky. (In der Realität kein einziger Kontakt unter 2 Meter länger als 5 Minuten.)

Bluetooth can pass through walls. Living near to a street or living next door to someone who is Covid-19 positive can be detected as a contact by the system although you were alone all the time in reality.

[PhilippBerl]

@neonhaxred Thank you, of course I have checked this possibility.Of course I have examined this possibility in detail. But none of my direct neighbors in our apartment building (all of them have CWA activated as well) have similar reports or are Covid positive.

Above all: In this case, the iOS exposure log would have to record two matches within the last 14 days? But there is no such match.

[neonhaxred]

Above all: In this case, the iOS exposure log would have to record two matches within the last 14 days? But there is no such match.

That's true! It might be this issue here: https://rp-online.de/panorama/coronavirus/corona-warn-app-fehler-in-ios-kann-zu-irrefuehrender-risikobewertung-fuehren_aid-53279797

[PhilippBerl]

The crucial question is still: is the report in CWA false positive (is the report of "2 Begegnungen mit niedrigem Risiko" wrong) or is the report in iOS false negative (is the report, that here haven't been any matches wrong)?

[neonhaxred]

It seems like nobody does know that. If you ask me I'd say that I rather trust the iOS protocol than the app itself.

[PhilippBerl]

I am not a programmer or software developer. But it seems logical to me that the decisive protocol is that of the iOS and the CWA evaluates the results of the protocol. My idea is that the iOS protocol lists the raw data that is modulated into a result in the app. My idea is, roughly speaking, that the iOS sends the information to the app. So if there are no matches in the exposure log, this rather indicates that there is a bug in the app?

Is there someone here who developed the app or who is responsible for developing the app and who can answer this question? It's actually about a real risk assessment.

[PhilippBerl]

And thank you very much for the answers!

But this question is still open to me and I think it is a crucial question to rely on the CWA App or not.

[Ein-Tim]

@PhilippBerl and @neonhaxred Please see #1106

The current assumption is that CWA is right, the ENF returns the correct numbers of Encounters to CWA but doesn't writes it in the Log.

I am living alone, spending 95% of my time alone in my flat. I had no personal contact in the last 14 days that would meet the conditions to be in any way risky. (In der Realität kein einziger Kontakt unter 2 Meter länger als 5 Minuten.)

@PhilippBerl regarding this part of your comment, please see #1234.

[Ein-Tim]

@GPclips

I think this issue can be closed since this problem seems to have been resolved by the release of iOS 13.7 / 14.

@GPclips could you close this Issue please? 🙂

[PhilippBerl]

@svengabr @GPclips @Ein-Tim @wenz To be honest, I do not quite understand why this issue was closed.

Has the problem been solved?

I don't know yet whether the two low-risk encounters with a covid-positive test person shown in the CWA actually took place, if so, when they took place (because there is no match in the exposure log) and whether it was the same person twice or two different ones.

I mention again that I was alone at home 95% of the time and personally had no contacts in the close range of <2 meters, which lasted longer than 5 minutes in the last weeks (the only one has been in the Staatsbibliothek zu Berlin to pick up books, in 5 minutes – exactly this was on 25.09. Except this situation I was only moving in my car, but without any personal contacts).

It can also be ruled out that I have a source of infection in a neighboring apartment, because the direct neighbors do not get an encounter reported on their CWA. I am in email/whatsapp-contact with all of our neighbors here in our apartment building, and most of them have the CWA.

I believe that my case can be helpful to solve the problem, because I had no personal contacts in the last weeks, except the one case described above on 09/25, at the same time no hits in the iOS exposure log. These are two remarkable features at the same time....

Is it possible that the Bluetooth signal in the courtyard 10 meters away would register a person tested positive for Covid and this would be reported as an encounter in the app?

Questions about questions.

[Ein-Tim]

@PhilippBerl

I don't know yet whether the two low-risk encounters with a covid-positive test person shown in the CWA actually took place, if so, when they took place (because there is no match in the exposure log) and whether it was the same person twice or two different ones.

But this Issue is "Exposure Checks only show 100 entries (checks of the last 7 days)", and this has been solved with iOS 13.7 and above. The Issue you are expieriencing is tracked here: #1106 (and will be solved with iOS 14.2)

And this:

I mention again that I was alone at home 95% of the time and personally had no contacts in the close range of <2 meters, which lasted longer than 5 minutes in the last weeks (the only one has been in the Staatsbibliothek zu Berlin to pick up books, in 5 minutes – exactly this was on 25.09. Except this situation I was only moving in my car, but without any personal contacts).

has been tracked here: #1234 (It's not 100% the same what you described but nearly the same.) Just if you're interested: Today I got another 2 Exposures with Low Risk on my secondary phone which was never outside my flat, and I also can't really belive it that these Encounters are true, but Since I'm on the iOS 14.2 Public Beta 2 on this phone and can see that the Logs in the Settings also say the same as CWA, I'm sure those Encounters happened and I'm also sure your encounters also happened. The deffinition of Low Risk Encounters can be found here: https://www.coronawarn.app/de/faq/#encounter_but_green

Now clearer why this Issue has been closed?

and whether it was the same person twice or two different ones.

I'm not quite sure about that, I think that you met 2 different persons. @ndegendogo am I correct?

Cheers 👍

[PhilippBerl]

Ok vielen Dank! Was sich auch nicht verstehe ist, warum wir hier alle auf Englisch miteinander kommunizieren, für eine App, die nur in Deutschland funktioniert. Aber offenbar soll alles etwas komplizierter sein als es unbedingt sein müsste ;-) , besonders, wenn es man besorgt ist and something is not working properly.

Many many Thanks for your kind answers. Nevertheless I still don't find my problem solved and hope to get more information.

I have posted the exposure log above. There are much more than just 100 entries reported and the protocol shows correctly the last 14 days.

Could please somebody give me more information?
I have now spent the last two days with this issue and have 10 neighbors in worry, asking me for explanation.

[Ein-Tim]

Wir schreiben hier auf Englisch um auch der Internationalen Community und vor allem Entwicklern aus anderen Ländern die Möglichkeit zu geben mitzuarbeiten. Wenn es dich nicht stört würde ich deine Fragen trotzdem kurz auf Deutsch beantworten 😀

I have posted the exposure log above. There are much more than just 100 entries reported and the protocol shows correctly the last 14 days

Ja, ich hab mir deinen Log angesehen und richtig, dort sind deutlich mehr als 100 Einträge und das ist auch gut so. Vor iOS 13.7 gab es den Fehler das in diesem Log nicht genug Einträge angezeigt wurden, deswegen gab es diesen Issue. Da Apple das aber ja mit iOS 13.7 gelöst hat, wurde der Issue jetzt geschlossen.

Wenn du sonst noch Fragen hast würde ich dich bitten Sie in #1106 zu stellen, dort könntest du auch noch deinen ersten Kommentar posten, wenn du magst. Aber musst du nicht, denn wie gesagt, dieser Fehler wird mit iOS 14.2 behoben.

[PhilippBerl]

@Ein-Tim Danke sehr! Thanks a lot.

I am quite sure that I am not alone with this problem and 100% sure, that there are many many people as I am, who are in worry, because they try to be as careful as they can to avoid any potential Corona-contact, but receiving the report in CWA that they had – accordingly to the app – a suspicious contact, or maybe more of them "Begegnung mit geringem Risiko"

And I am very very sure that more than 80% of these people, insecure and frightened do not know what to do. I assume most of them even would not have enough patience, time and maybe energy and technical knowhow to find this site here

[PhilippBerl]

So, @Ein-Tim, do you think the two contacts with low risk ("Begegnungen mit geringem Risiko") are real int the meaning, that they have happened? Thank you very much, once again!

[Ein-Tim]

@PhilippBerl Yes, I'm very sure these contacts happend, and even if this was someody walking on the other side of the street or something like this, the App will show it to you.

And I am very very sure that more than 80% of these people, insecure and frightened do not know what to do. I assume most of them even would not have enough patience, time and maybe energy and technical knowhow to find this site here

I (and also the devs here) know that these Low Risk Encounters are a little bit tricky to understand, but thats why with version 1.3.2 you should see some more details about your Low Risk Encounter in the App.

[PhilippBerl]

@Ein-Tim Again, thank you very, very, very much for your kind and helpful answers.

I have just received the 24h-updated report from my CWA, still reporting the contacts with low risk.

If the contacts really happened, my only explanation indeed would be, that I walked around the street or just sitting in the car in the traffic jam and somebody farer away in the meantime has been tested positive. But in this cases the contact would have been <5 minutes and the distance >8 meters (in normal distance, but the Bluetooth-signal isn't that precise, if the iPhone has been in my jacket or in my car or in my apartment etc.) I agree totally with you, that these cases would be the only ones to explain, why I have to suspicious contacts reported in my CWA. Once again: Is a short contact <2-5 minutes and farer than 8 meters away really already registered as contact? I have read that these cases are filtered out, so that they are not reported in any case.

I have updated the CWA yesterday to 10.3.2., before that I had the first version something like 10.01. or so, but I still have no more information about these two matches. The situation in the App is the same as yesterday. I have also updated my iOS to 10.14.01. today without any changes regarding the exposure log, but this seems to be evident to me, unfortunately

[Ein-Tim]

Again, thank you very, very, very much for your kind and helpful answers.

No problem! @PhilippBerl

I have just received the 24h-updated report from my CWA, still reporting the contacts with low risk.

Means the contacts aren't already 14 days in the past, 14 days after the Encounter the App won't show them anymore.

Once again: Is a short contact <2-5 minutes and farer than 8 meters away really already registered as contact?

Yes they are, see @kbobrowski's comment here: https://github.com/corona-warn-app/cwa-website/issues/307#issuecomment-679233521

I have read that these cases are filtered out, so that they are not reported in any case.

Thats not true. Every Encounter will be shown to the User. (In earlier Versions of the FAQ there was the statement that some encounters are "verworfen", but since this isn't true, it was corrected.)

I have updated the CWA yesterday to 10.3.2., before that I had the first version something like 10.01. or so, but I still have no more information about these two matches. The situation in the App is the same as yesterday.

Nice to hear that you updated, just a quick hint from me: I don't know how you do it with other Apps but I would recommend to keep CWA always up to date 👍 Regarding no more Information: If you click on the Green Card, this Screen should open:

There you can find more Information about Low Risk Encounters.

I have also updated my iOS to 10.14.01. today without any changes regarding the exposure log, but this seems to be evident to me, unfortunately

Apple didn't change the Log with iOS 14.0.1, but they will solve the Issue #1106 with iOS 14.2

[PhilippBerl]

@Ein-Tim Thank you, great!

I have read that these cases are filtered out, so that they are not reported in any case.

Thats not true. Every Encounter will be shown to the User. (In earlier Versions of the FAQ there was the statement that some encounters are "verworfen", but since this isn't true, it was corrected.)

This was one of the answers I was searching for.

It was this formulation, which is most irritating "is discarded" ("wird verworfen") I would like to strongly recommend the developers or maintainers of the CWA to clarify this explicitly somewhere, because this explanatory text can be accessed unchanged and seems to be actual.

If you click on the Green Card, this Screen should open: Thanks also for this hint. I already know this page of the CWA, but it is of course much too general for a scientist like me and someone who wants to know the source of a danger exactly.

Thank you also very much for that information. I already knew this Screen, but it is of course much too general for a scientist like me and someone who wants to know the source of a risk exactly, although I respect the problem of Privacy Policy (Datenschutz).

So once again a question of understanding. Assuming that the exposure log protocol in iOS would work without problems, these two matches would be displayed with date. That means: They would be completely anonymous, but at least I would know when the encounter was. And then I could better assess my personal risk. If, for example,CWA shows a contact with low risk, than I can see into the log and find the match. Let was say it was x-day. If I have been only at home that day, I would know that someone was maybe walking through the courtyard who later had been tested positive; or I would know "Aha, there was a parcel carrier who left a package at my door" or, I had fetched my books from the library etc. This is what I miss at the moment. Am right in assuming, that if everything would have worked properly with iOS , this would be a way to better limit the risk encounter and thus better assess the actual risk individually? And that's where I still see the problem: As I said, in my case it is an extreme coincidence that the app shows these encounters and that they are not reported in the exposure log.

[daimpi]

Assuming that the exposure log protocol in iOS would work without problems, these two matches would be displayed with date. That means: They would be completely anonymous, but at least I would know when the encounter was. And then I could better assess my personal risk. If, for example,CWA shows a contact with low risk, than I can see into the log and find the match. Let was say it was x-day. If I have been only at home that day, I would know that someone was maybe walking through the courtyard who later had been tested positive; or I would know "Aha, there was a parcel carrier who left a package at my door" or, I had fetched my books from the library etc. This is what I miss at the moment. Am right in assuming, that if everything would have worked properly with iOS , this would be a way to better limit the risk encounter and thus better assess the actual risk individually?

Unfortunately not. Even if EN log was working properly there is currently no direct way to find out when the green encounter happened (unless you have a rooted Android phone, in which case you can use corona-warn-companion-android to get more info). There is only an indirect way to find out when the encounter was, by checking when the encounter disappears in CWA. I've written about this in some more detail here if you're interested.

I also wish there was more information shown for green encounters and we have an issue for this which you can upvote here 🙂.

[Ein-Tim]

This is how it would look like if everything would be correct...

[PhilippBerl]

@Ein-Tim Exactly! And mine looks like this:

Now, in the meantime I have asked my neighbor to send me her protocol. Her CWA shows 1 "Begegnung mit geringem Risiko". Her protocol also shows no match in the last 14 days BUT her protocol shows several matches in 15.09. and 16.09. and INDEED she had been at a funeral with 20 persons, of which 15 (!) later tested positive. I don't want to know if her CWA signaled red, but she had been tested negative and has been healthy ever since. But I try to understand: Are these matches logged, precisely because they were in fact very likely to be classified as risky encounters?

In any case, it proves that it makes sense to use the protocol to narrow down the date of the matches also in case of "Begegnungen mit geringen Risiko". Or is there an error in reasoning? At least it is the date on which the server sends the data to the iPhone, and in this way it is possible to narrow things down.

P.S. Her iOS is 10.3.7 and CWA is updated

[PhilippBerl]

@daimpi Unfortunately not. Even if EN log was working properly there is currently no direct way to find out when the green encounter happened (unless you have a rooted Android phone, in which case you can use corona-warn-companion-android to get more info). There is only an indirect way to find out when the encounter was, by checking when the encounter disappears in CWA. I've written about this in some more detail here if you're interested.

But I don't understand that, please help me to understand it. If EN (EN is the acronym for the "protocol", I assume) shows 1 Match as @Ein-Tim has just shown it, than the match is related to a specific date. And that date gives at least an idea, when the contact happened.

[Ein-Tim]

@PhilippBerl Yeah, your log is showing no match because you are not on the iOS 14.2 Public Beta 2 as I am.

But I try to understand: Are these matches logged, precisely because they were in fact very likely to be classified as risky encounters?

I don't know when something is still loged and when not, but normally everything should be loged.

In any case, it proves that it makes sense to use the protocol to narrow down the date of the matches also in case of "Begegnungen mit geringen Risiko". Or is there an error in reasoning? At least it is the date on which the server sends the data to the iPhone, and in this way it is possible to narrow things down.

Yeah maybe the log can help you to narrow it down but to know when a Low Risk Encounter occured you just have to take a look in the App every day and when the Encounter isn't shown anymore, then you know your encounter happened 14 days ago. You understand? 😀

P.S.: EN is the abbreviation for Exposure Notification.

than the match is related to a specific date

Yes, but the time shown in the Settings is only the Time when your App downloaded the keys from the Server, not the Time you met someone.

[PhilippBerl]

Great, @Ein-Tim After 24 hours I really start to understand how the CWA works. Amazing system. Although it is obviously the best compromise between Datenschutz and Infektionsschutz (sorry for writing this in German), there is really place to improve CWA by allowing either EN or CWA or both, to help the consumer and potential infected person to narrow down the date of the suspicious contact. This really would be a great help, in particular regarding the fact that unfortunately many people don't use the CWA. This would help to improve the personal risk assessment.