GraphQL: error "The requested service is temporarily unavailable." on some queries

[Volkhard_Vogeler]

Hello,
we are using the Cloud-Instance on realm. On some GraphQL query requests we get the error message “The requested service is temporarily unavailable.” with the hint “https://docs.realm.io/server/troubleshoot/errors#service-unavailable”.
In the Realm-Logs there is an entry “Realm Object Server has started and is listening on http://0.0.0.0:9080” then - it seems that there is a new instance startet.

Following Questions for this:

1. what are the possible reasons for this?
2. when i follow the link “https://docs.realm.io/server/troubleshoot/errors#service-unavailable” i cant´t see the content - there is a Login screen displayed. even if i Register i can´t ee the content. Why and What is the content?

best regards

Volkhard

[Stennie]

Hi Volkhard,

The documentation link you found is for the on-premise version of Realm Object Server, which is no longer distributed for new installations (or relevant for troubleshooting this issue on Realm Cloud).

The “503: Service Temporarily Unavailable” message is a general error code that will need investigation. For example, this may be a transient issue, something specific to your instance, or perhaps the result of excessive resource consumption for some GraphQL queries.

Please create a support case for the team to investigate. It would be helpful if you can include an example of a query and date/time when the issue was observed.

Thanks,
Stennie

[Etienne_Hanser]

Same error here… Support isn’t helping. We are about to go productive. Are you guys serious?

[Stennie]

Hi @Etienne_Hanser,

Reviewing your support case, I can see the team started investigating and corresponding the same day you filed your issue.

I appreciate your frustration in the time to resolve this issue, but investigation of resource contention for shared infrastructure can take longer than dedicated deployments. GraphQL queries can be resource intensive, but the impact depends on your specific queries and workload.

I noticed there were some adjustments to increase allocated resources for your cluster, but this wasn’t communicated to you as clearly as it could have been. When the team member was asking if your issue was resolved, they should have mentioned that they wanted to confirm the results of some resource adjustments they were monitoring.

If you are preparing to go live with a production workload, I’d also suggest doing come capacity planning and consider if you have the right plan for your requirements. The Realm Cloud Standard plan ($30/month) is based on multi-tenant/shared resources, so performance is less predictable and scalable than a dedicated deployment.

The Standard plan also does not include an SLA or extended support coverage, so you need to have realistic expectations around support response time. We have been ramping quite a few new team members and the support experience should be significantly improving.

However, we definitely appreciate any feedback on how we can improve.

Regards,
Stennie