id | slug | title | image | description | date | tags | ||||
---|---|---|---|---|---|---|---|---|---|---|
kibDataPlugin |
/kibana-dev-docs/key-concepts/data-plugin |
Data services |
The data plugin contains services for searching, querying and filtering. |
2020-12-02 |
|
The data plugin provides common data access services, such as search
and query
, for solutions and application developers.
The autocomplete service provides suggestions for field names and values.
It is wired into the TopNavMenu
component, but can be used independently.
The getQuerySuggestions
function helps to construct a query.
// `inputValue` is the user input
const querySuggestions = await autocomplete.getQuerySuggestions({
language: 'kuery',
indexPatterns: [indexPattern],
query: inputValue,
});
The getValueSuggestions
function returns suggestions for field values.
This is helpful when you want to provide a user with options, for example when constructing a filter.
// `inputValue` is the user input
const valueSuggestions = await autocomplete.getValueSuggestions({
indexPattern,
field,
query: inputValue,
});
The query service is responsible for managing the configuration of a search query (QueryState
): filters, time range, query string, and settings such as the auto refresh behavior and saved queries.
It contains sub-services for each of those configurations:
data.query.filterManager
- Manages thefilters
component of aQueryState
. The global filter state (filters that are persisted between applications) are owned by this service.data.query.timefilter
- Responsible for the time range filter and the auto refresh behavior settings.data.query.queryString
- Responsible for the query string and query language settings.data.query.savedQueries
- Responsible for persisting aQueryState
into aSavedObject
, so it can be restored and used by other applications.
Any changes to the QueryState
are published on the data.query.state$
, which is useful when wanting to persist global state or run a search upon data changes.
A simple use case is:
function searchOnChange(indexPattern: IndexPattern, aggConfigs: AggConfigs) {
data.query.state$.subscribe(() => {
// Constuct the query portion of the search request
const query = data.query.getEsQuery(indexPattern);
// Construct a request
const request = {
params: {
index: indexPattern.title,
body: {
aggs: aggConfigs.toDsl(),
query,
},
},
};
// Search with the `data.query` config
const search$ = data.search.search(request);
...
});
}
You can also retrieve a snapshot of the whole QueryState
by using data.query.getState()
data.query.timefilter
is responsible for the time range filter and the auto refresh behavior settings.
Timefilter provides an API for setting and getting current auto refresh state:
const { pause, value } = data.query.timefilter.timefilter.getRefreshInterval();
data.query.timefilter.timefilter.setRefreshInterval({ pause: false, value: 5000 }); // start auto refresh with 5 seconds interval
Timefilter API also provides an autoRefreshFetch$
observables that apps should use to get notified
when it is time to refresh data because of auto refresh.
This API expects apps to confirm when they are done with reloading the data.
The confirmation mechanism is needed to prevent excessive queue of fetches.
import { refetchData } from '../my-app'
const autoRefreshFetch$ = data.query.timefilter.timefilter.getAutoRefreshFetch$()
autoRefreshFetch$.subscribe((done) => {
try {
await refetchData();
} finally {
// confirm that data fetching was finished
done();
}
})
function unmount() {
// don't forget to unsubscribe when leaving the app
autoRefreshFetch$.unsubscribe()
}
Provides access to Elasticsearch using the high-level SearchSource
API or low-level Search Strategies
.
The SearchSource
API is a convenient way to construct and run an Elasticsearch search query.
const searchSource = await data.search.searchSource.create();
const searchResponse = await searchSource
.setParent(undefined)
.setField('index', indexPattern)
.setField('filter', filters)
.fetch();
One benefit of using the low-level search API, is partial response support, allowing for a better and more responsive user experience.
import { isRunningResponse } from '../plugins/data/public';
const search$ = data.search.search(request)
.subscribe({
next: (response) => {
if (!isRunningResponse(response)) {
// Final result
search$.unsubscribe();
} else {
// Partial result - you can update the UI, but data is still loading
}
},
error: (e: Error) => {
// Show customized toast notifications.
// You may choose to handle errors differently if you prefer.
data.search.showError(e);
},
});