1.This PR adds the output format to the external destination either json
/ ndjson , bydefault it is json
---------
Co-authored-by: oasisk <ashish.j.kolhe@gmail.com>
1. This PR fixes the log page download count mismatch issue so when user
tries to get 1 to 100 records we are actually sending from as 1 which is
incorrect because BE expects from 0 so modified we now send -1 from user
input eg: user enters 1 we send it as 0 , user sends 50 send 49 this
will make sure that actual records are being downloaded
2. This PR also fixes the service account issue so we dont need first
name and last name for the service account so we removed that and added
description in place of that , removed them in the service account list
as well |
#6887#6837
This histogram metric should allow understanding the age of files
handled by querier.
Example queries
`histogram_quantile(0.95,
rate(zo_file_access_time_bucket{stream_type='logs'}[5m]))`
---------
Co-authored-by: Hengfei Yang <hengfei.yang@gmail.com>
This PR fixed the issue when compactor not working for a long time, it
will have millions files need to be update stream stats. it will take a
long time or time out.
- Verify error handling and no results found with histogram
- Verify error handling with histogram toggle off and on
- Verify histogram toggle persistence after multiple queries
- Verify histogram toggle with empty query
- Verify histogram toggle with complex query
- Parser fails to parse Data fusion functions which throws error.
- Pass original user written query if not able to parse.
Co-authored-by: Hengfei Yang <hengfei.yang@gmail.com>
Fixes#6928
Currently if the root user resets the ingestion passcode token, the
cache is not updated for root user, hence on refresh it still shows the
old passcode even though it is changed. As a result, the ingestion
passcode token shown on UI does not work.
impl #6948
- [x] When we run query, we get node list and check the node CPU, drop
at most 20% querier nodes.
- [x] When download file, we first try to download from original node(
by consistent hashing), then download from object store.
- [ ] ? should also apply this logic for ingester node ?
- [x] Fixed metrics keep increase issue of DownloadManager
Fixes permission check for organization related api. In ofga when we set
a role with org perms, it sets to `org:_all_{org}` but in the check , we
check for permission on `org:{org}` which causes users with role having
permission on org, not be able to rename org. This fixes it by updating
it to correct check.
---------
Signed-off-by: Yashodhan Joshi <yjdoc2@gmail.com>
Co-authored-by: Yashodhan Joshi <yjdoc2@gmail.com>
the values response hits always honor the size, but the total field
which indicates the count of responses is returning incorrect results.
This PR fixes it.
This PR also fixes a migration issue where the ID field were used as i64
ref: #5612
We see an overflow when we multiply the query duration to a large
storage size, to fix it we divide the storage size first with the data
retention time
this PR introduces o2 ai agent with this user able to ask o2 ai about
sql query , vrl query and logs
---------
Co-authored-by: Subhradeep Chakraborty <chakrabortysubhradeep556@gmail.com>
Co-authored-by: Taiming Liu <liutaiming3@gmail.com>
The issue is when we get cluster list we didn't filter the alert querier
cluster then the stats duplicated. The fix in o2 enterprise. this pr
only update some logs.
In dashboard panel edit mode, repeatedly adding and closing a drill-down
with an SQL query causes the UI to freeze due to a memory leak in the
Quasar dialog component.
Co-authored-by: ktx-vaidehi <vaidehi.akhani@kiara.tech>
Hengfei Yang
Sai Nikhilandoasisk
Taiming LiuandSubhra264
udevandHengfei Yang
Omkar Kesarkhane
Neha P
Abhay PadamaniandHengfei Yang
Shyam PanjiyarandHengfei Yang
ktx-akshay
Siddharth
Bhargav
Ashish Kolhe
Yashodhan JoshiandYashodhan Joshi
Huaijinandhengfei yang
Patrick Sullivan
