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
This ticket has been raised on reports of HTTP 429 errors within PD Live.
It has been noted that some users are trying to pull large historical datasets which PD Live was not designed/intended for.
Specifically, this occurs under the following user behaviours:
Query by long since date (i.e. too many open incidents)
Query by resolved incidents (i.e. show me everything)
The current implementation of PD Live will initially fetch incident data in parallel, then do batched calls for the incident notes.
When users decide to update the query upon mid-fetch, then additional API calls are made which exceed the rate limit, ultimately leading to HTTP 429 errors and empty/missing data in the UI.
Proposed Change
To prevent HTTP 429 errors, we have proposed the following:
Include debounce on query update (see entrypoint below)
i.e. only query when you have completing updating the remaining filters.
Summary
This ticket has been raised on reports of
HTTP 429
errors within PD Live.It has been noted that some users are trying to pull large historical datasets which PD Live was not designed/intended for.
Specifically, this occurs under the following user behaviours:
The current implementation of PD Live will initially fetch incident data in parallel, then do batched calls for the incident notes.
When users decide to update the query upon mid-fetch, then additional API calls are made which exceed the rate limit, ultimately leading to HTTP 429 errors and empty/missing data in the UI.
Proposed Change
To prevent HTTP 429 errors, we have proposed the following:
Include debounce on query update (see entrypoint below)
i.e. only query when you have completing updating the remaining filters.
pd-live-react/src/redux/incidents/sagas.js
Line 79 in 68991d6
Find total incidents for query and prompt if over specified limit of
200 incidents
(see [Enhancement] Prompt user if incident query returns a large dataset #52)
Implementing these changes should also address #77 and #92 given the fixed number of incidents in view.
The text was updated successfully, but these errors were encountered: