INP (still) not showing on self hosted #3322
Comments
|
🤔 |
|
I also added console output is (after clicking the button and switching the tab) |
You'll need to re-run the But since you're on 24.8.0, I guess you'll need to match your flags to what's on that release, see here: self-hosted/sentry/sentry.conf.example.py Lines 263 to 326 in b7ed3cb A few things that are missing from yours: "projects:span-metrics-extraction-addons",
"organizations:standalone-span-ingestion",
"organizations:indexed-spans-extraction",I'd suggest to update the flags to match the one on the 24.8.0 tag first, then re-run |
|
@aldy505 thank you - I will try as soon as possible. But self-hosted/sentry/sentry.conf.example.py Lines 263 to 326 in b7ed3cb |
getsantry
bot
added
Waiting for: Product Owner
and removed
Waiting for: Community
labels
|
AFAIK,
I usually hover over https://github.com/getsentry/sentry/blob/master/src/sentry/features/temporary.py and see the commit history of that specific file. That leads somewhere..... usually. |
|
Still not working :-( I re-ran install.sh and triggered sending INP values via my inp.html again Still: "No INP data found in this project":
And INP seems to be sent as per network tab:
|
getsantry
bot
added
Waiting for: Product Owner
and removed
Waiting for: Community
labels
|
Wondering if this is in any way related to this similar issue which happened on the SaaS. Unfortunately there is no information on what was fixed in the SaaS exactly. |
|
INP usually take a while. You might want to come back in a few hours (or maybe a few days) after sending lots of INP data. |
|
But why is that? Does it have to do with sampling? Because my local sample rate is 1.0 Is there additional sampling happening? How much data is needed to calculate the value? Cause I am testing that manually currently and would have to press that button 100 times then ;-) ? I do have it running on prod, too though (but with 0.1 sample rate and an older SDK (https://browser.sentry-cdn.com/8.7.0/bundle.tracing.min.js) - so there is still hope, something will show up after the weekend... EDIT: And I would at least expect the INP to appear instantly in the event-line, because all the other values (LCP, FCP) appear there already. |
getsantry
bot
added
Waiting for: Product Owner
and removed
Waiting for: Community
labels
I have no idea on this, I usually don't expect Sentry stuff to appear instantly other than reported errors and alerts. |
|
Yeah, I also noticed some kind of delay for the page views to appear together with the LCP + FCP data (maybe ~ 1 minute), but when it appears I would expect INP to be also filled. I guess it's still not working, because I just rechecked and the INP is still empty. With a little bit help from ChatGPT I dug into the clickhouse database: # docker exec -it sentry-self-hosted-clickhouse-1 clickhouse-client
SELECT
project_id,
transaction_name AS page,
start_ts AS start_timestamp,
finish_ts AS finish_timestamp,
arrayJoin(measurements.key) AS measurement_key,
arrayJoin(measurements.value) AS measurement_value
FROM
transactions_local
WHERE
transaction_name = '/inp.html'
AND arrayJoin(measurements.key) = 'inp'
ORDER BY
start_ts
FORMAT Vertical;This does not yield results. Does it for you? |
getsantry
bot
added
Waiting for: Product Owner
and removed
Waiting for: Community
labels
@amenk This is what I got (removed the SELECT
project_id,
transaction_name AS page,
start_ts AS start_timestamp,
finish_ts AS finish_timestamp,
arrayJoin(measurements.key) AS measurement_key,
arrayJoin(measurements.value) AS measurement_value
FROM transactions_local
WHERE arrayJoin(measurements.key) = 'inp'
ORDER BY start_ts ASC
FORMAT Vertical
Query id: 38d8b597-d54f-4d3f-a99e-62733d557a7b
Ok.
0 rows in set. Elapsed: 0.393 sec. Processed 1.62 million rows, 56.72 MB (4.12 million rows/s., 144.22 MB/s.)
Peak memory usage: 21.17 MiB.I don't think that's the right query for it. |
|
Yeah, the query is wrong - I also don't see data, now, even I have now data in the dashboard.
Also I can see the element attribution well, when going to interactions:
For the record, I also added the flag But in the end of the day, it just seems to need some good amount of data to be displayed. My local tests with /inp.html still don't display any data, but the production values are shown, now. It would be interesting to know, how much data is actually necessary for this to be displayed, in case anyone is testing this :-) Anyways, thank you every much @aldy505 for your help and I will close this issue. |
Self-Hosted Version
24.8.0
CPU Architecture
x64_64
Docker Version
27.1.2, build d01f264
Docker Compose Version
2.29.1
Steps to Reproduce
Follow up of
getsentry/sentry-javascript#13666 & #2970
I am hosting the test file from getsentry/sentry-javascript#13666 (comment) locally via Apache with a DSN set to my self hosted instance.
Request with INP data are showing in the network panel and get a "200 ok", but nothing shows up in sentry
sentry.conf.py feature flags:
Expected Result
INP data is showing.
Actual Result
Event ID
fafed2c4858f46ea8114f133bafa79cb
The text was updated successfully, but these errors were encountered: