Hello ESRI Community,
I am currently experiencing performance issues when accessing and updating hosted feature layers via the REST API.
Specifically, the services are hosted at:
https://services9.arcgis.com/3wBJQO6LK1gVuxyq/
The issues I am encountering:
Delays and frequent 504 Gateway Timeout errors when performing update operations (applyEdits)
Inconsistent performance (sometimes fast, sometimes very slow without a clear pattern)
Health dashboards https://trust.arcgis.com/en/system-status/ indicates no ongoing issues, despite the lag I am experiencing
Any insights, suggestions, or similar experiences would be greatly appreciated.
Solved! Go to Solution.
Hey Stad,
This can be a challenging area, technically at times like finding a needle in a haystack.
Is there anything observed that might rule where the issue is coming from? I'm thinking Client-side issues such as hardware struggling to render, service level specific issues, webmap/app level issues, or an overall performance issue on your ArcGIS Online organisation?
For starting points I would take a look at your ArcGIS Online Health Dashboard, and ensure we are not seeing a prolonged high average usage, as ArcGIS Online will maintain a lot of health measures where requests are throttled if there is a risk of Datastore activity causing a crash - This guide here is useful for troubleshooting at your Organisational datastore level.
If things look fine at the datastore level, then I would recommend comparing your Web Maps and services. Are these all running slow, or specific ones? If you do a ctrl - f5 refresh of the Web Map/App while using the web browser dev tools, and check the network tab how long are the slowest requests taking?
It can be a subjective area, as some configurations need a lot of requests, and are fine with a 2 second query time, while others may find that unacceptable, but the key thing I'd look out for is if queries are finishing wthin this timeframe, or if they get slower and slower, occasionally on some web map configurations we can see queries take over 20 seconds each.
If you think the issue is at this level, then I would try reduce the number of requests being sent, using Web Map optimisations such as clustering, scale thresholds, aligning Basemap and Layer projections etc.
Service level optimisations may also help you, such as trimming change logs, and performing a spatial reindex if services are fairly old, and have been edited.
Hopefully that gives you a few pointers. As George said above don't be afraid of contacting support as this can be a tricky area to troubleshoot.
David
I would log a case with Technical Support to investigate this issue.......
Hey Stad,
This can be a challenging area, technically at times like finding a needle in a haystack.
Is there anything observed that might rule where the issue is coming from? I'm thinking Client-side issues such as hardware struggling to render, service level specific issues, webmap/app level issues, or an overall performance issue on your ArcGIS Online organisation?
For starting points I would take a look at your ArcGIS Online Health Dashboard, and ensure we are not seeing a prolonged high average usage, as ArcGIS Online will maintain a lot of health measures where requests are throttled if there is a risk of Datastore activity causing a crash - This guide here is useful for troubleshooting at your Organisational datastore level.
If things look fine at the datastore level, then I would recommend comparing your Web Maps and services. Are these all running slow, or specific ones? If you do a ctrl - f5 refresh of the Web Map/App while using the web browser dev tools, and check the network tab how long are the slowest requests taking?
It can be a subjective area, as some configurations need a lot of requests, and are fine with a 2 second query time, while others may find that unacceptable, but the key thing I'd look out for is if queries are finishing wthin this timeframe, or if they get slower and slower, occasionally on some web map configurations we can see queries take over 20 seconds each.
If you think the issue is at this level, then I would try reduce the number of requests being sent, using Web Map optimisations such as clustering, scale thresholds, aligning Basemap and Layer projections etc.
Service level optimisations may also help you, such as trimming change logs, and performing a spatial reindex if services are fairly old, and have been edited.
Hopefully that gives you a few pointers. As George said above don't be afraid of contacting support as this can be a tricky area to troubleshoot.
David
Are you able to even view any of the services? If I click your link, it opens the rest/services page, but if I try to open any of the feature services, it eventually times out.
Not familiar with the trust.arcgis.com health dashboard, but it looks to be similar to this one.
Past experience has shown that the Health dashboard just as well be a static image as it will normally show green, even when the servers are down completely. With luck, it will update a few days later and show the problems in the resolved in the last 30 days section. But, even then, not likely.
I would suggest you log a tech support case on this so they can get the server working again.
R_
This issue is currently being investigated. Thank you all for your assistance.