1. Remove action_id from Histogram and Page-count search payload
2. Rename Action type `Service` to `Real Time`
3. Added validation for service account selection while creating action
This PR includes changes
1. introduced icons for histogram , sql mode, quickmode and moved search
history , download logs , search scheduled jobs options into menu that
will be there at right end
2. this will make the UI to accommodate upto 1200px with datetime as
relative , 1350 when it is absolute
3. introducing icons saved space of ~200px in total for both sides
implemented alerts folders in the UI
1. search across folders
2. alert listing page moved condition and description column to the
expand so whenever user clicks on any row it will show respective
condition and description if there
3. provided fixed width for columns so that all columns are in view port
visibility expect name column as it can be varied depending upon name
given by user
4. search across all folders implemented in this PR only
Import alert add ons
1 additional things done to import alert as well like introducing the
folders in import page
---------
Co-authored-by: omkark06 <omkar@zinclabs.io>
Co-authored-by: Subhradeep Chakraborty <chakrabortysubhradeep556@gmail.com>
- Fixes 403 error when alert move api is called
- 403 error when trying to create alert in default folder even after
having permission
- Fixes # Regular user without `GET` user permission should not be able
to get list of all users
Currently in the main, when creating a dashboard, we check if the
`owner` field is there in the backend. If it is not present, it returns
Missing Owner error. This PR fixes this by checking first if the `owner`
field is non-empty, otherwise it uses the currently logged in user as
`owner`.
When a pipeline takes more than the scheduler timeout to complete the
task, the scheduler background job `watch_timeout` increases the
`retries` field by 1 and again updates the scheduled jobs status back to
`waiting` from `processing`. If this happens three times, the `retries`
field value go beyond the max retries and hence is never pulled by
scheduler again.
With this change, whenever we process a derived stream we first check if
the current `retries` field matches with the max retries. If it matches
then we simply make the `retries` `0` and again run it after the
scheduled frequency. The start_time for the query is not changed though,
so there should not be any data loss when quering.
This PR fixes couple of issues
1. changed the warning message to minimum frequency should be -- minutes
when user tries to give less than minimum frequency interval in alert
and pipeline
2. other is made period in sync with frequency so whenever user toggles
the cron so in pipelines we are auto selecting the period according the
frequency so toggling not picking up the latest values of freq to period
The search response should be cached only when it is not a partial error
or if it is so, only for range errors we should cache the result. This
is a hacky fix which caches the result when the above condition is
satisfied.
Partly addresses #6311 (does not fix the root cause where is_partial is
false even when there is function error).
---------
Co-authored-by: oasisk <ashish.j.kolhe@gmail.com>
this PR will introduce a validation check that if the query is
validating then we need to disable the validate and close button so that
user don't click on it multiple times and also changed text from
validate and close to validating while the query is validating.
fix: #6120
2 adjustments applied to pipeline queries comparing to normal queries,
when the stream has `store_original` set to true
- [x] stop automatically skipping `_original` for `select * ...` queries
- [x] stop automatically adding `_o2_id` when not selected
With these adjustments, users would need to explicitly select those two
fields when needed.
- Fixes#6289
- Also, Removes use of semaphore inside the scheduler. Currently it only
uses the mpsc channel and waits if the mpsc channel is full. Apart from
this, it uses db locks in postgres and mysql when pulling
scheduled_jobs.
This PR implement use querier node as query leader for AlertManager.
> Only for single stream query(standard search request), there is no
gRPC for MultipleStreamQuery request for now.
Taiming Liu
Subhra264
Shyam Panjiyar
Omkar Kesarkhane
udev
ktx-vaidehi
Hengfei Yang
Abhay Padamani
Zhao.wei
Ashish Kolhe
Patrick SullivanandHengfei Yang
dependabot[bot]