You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The indexer currently subscribes (backed by registry webhook) to the dataset changes event stream and auto-indexing any dataset changes.
This process may introduce a few seconds delay between the changes made on the original dataset in the registry and the changes indexed in the search engine.
The short time delay won't cause too much trouble for general-purpose updates. But for access control related updates, we will want a proactive interface to index the changes without delay.
This ticket is about adding 2 new APIs to the indexer:
index dataset by id: PUT /v0/dataset/:id
indexer will attempt to retrieve the dataset data from the registry using the ID and index the new data into search engine
delete dataset by id: DELETE /v0/dataset/:id
indexer will delete the dataset with the specified id from the search engine
Once the APIs are added, we also need to expose them via gateway & add auth to existing APIs (previously are internally only APIs). For existing APIs, we should:
exposing the following APIs.
POST /reindex : trigger a proactive full index.
validate auth via operationUri api/indexer/reindex.
POST /reindex/snapshot: trigger a snapshot.
validate auth via operationUri api/indexer/reindex/snapshot.
We don't use this API at this moment as it might not fully functional due to this ticket
GET /reindex/in-progress: get progress info of the full index trigger.
validate auth via operationUri api/indexer/reindex/in-progress.
DELETE /dataset/{datasetId}: delete a dataset from search engine index
validate auth via operationUri object/dataset/delete
PUT /dataset/{datasetId}: reindex a single dataset
validate auth via operationUri object/dataset/update
The following APIs will still stay internal only.
POST /registry-hook: web-hook listener
stay as an internal API for performance consideration
Status API e.g. /status/ready & /status/live
stay as internal API as there is for internal usage only
The text was updated successfully, but these errors were encountered:
Indexer: Adhoc dataset indexing API
The indexer currently subscribes (backed by registry webhook) to the dataset changes event stream and auto-indexing any dataset changes.
This process may introduce a few seconds delay between the changes made on the original dataset in the registry and the changes indexed in the search engine.
The short time delay won't cause too much trouble for general-purpose updates. But for access control related updates, we will want a proactive interface to index the changes without delay.
This ticket is about adding 2 new APIs to the indexer:
/reindex
: trigger a proactive full index.api/indexer/reindex
./reindex/snapshot
: trigger a snapshot.api/indexer/reindex/snapshot
./reindex/in-progress
: get progress info of the full index trigger.api/indexer/reindex/in-progress
./dataset/{datasetId}
: delete a dataset from search engine indexobject/dataset/delete
/dataset/{datasetId}
: reindex a single datasetobject/dataset/update
/registry-hook
: web-hook listener/status/ready
&/status/live
The text was updated successfully, but these errors were encountered: