Unplanned reboot of API Management node

Ankit Raj 1

At Wednesday, September 21, 2022 at 6:20:02 AM GMT+5:30, the Azure monitoring system received the following information regarding your API Management service:
The compute node that powers your API Management instance is being rebooted. A reboot of the API Management's underlying compute node is most likely triggered by an automatic upgrade.

There is also vnet connection issue at my end:

Failed to connect to management endpoint at xxxxxxxxx.management.azure-api.net:3443 for a service deployed in a virtual network. Make sure to follow guidance at https://aka.ms/apim-vnet-common-issues.

The SLA is DEV so we have not any SLA provided,

Connection to management endpoint failed with WebException: ConnectFailure : Unable to connect to the remote server

1 answer

JananiRamesh-MSFT 21,861 Reputation points

2022-09-21T18:39:37.243+00:00

Hi @Ankit Raj Thanks for reaching out. As you mentioned that this is a Developer SKU service, it has one machine, so when it runs to a service update/upgrade or OS upgrade, the machine will go down and therefore, all the service components will go down as well.

Is the issue being still existing? As you mentioned that your service is deployed in a VNET could you please verify if the configurations are made as per the document https://learn.microsoft.com/en-us/azure/api-management/api-management-using-with-vnet?tabs=stv2#configure-nsg-rules.

Dev SKU doesn't offer SLA as it features a single VM you can always upgrade to higher SKUs for better availability (other tiers have SLA because of machine redundancy, in pairs).

reference: https://azure.microsoft.com/en-us/pricing/details/api-management/

let me know incase of further queries, I would be happy to assist you.

Please 'Accept as answer' and ‘Upvote’ if it helped so that it can help others in the community looking for help on similar topics.
Please sign in to rate this answer.
Ankit Raj 1 Reputation point

2022-09-26T05:31:56.037+00:00

Hi @JananiRamesh-MSFT , issue has been resolved. but now we need to change the plan of the api instance for the 100 % uptime. Do we face the same issue in basic tier as well? And can we setup any alert mechanism that can notify us when api management instance is down?.

JananiRamesh-MSFT 21,861 Reputation points

2022-09-26T05:58:46.217+00:00

Hi @Ankit Raj , Unfortunately, service upgrade/OS upgrade is something that you cannot control. The VM that is underlying for that service needs sometime to receive a mandatory update. For basic tier you will get 2VMs so the complete service will not go down but no 100 % uptime guaranteed please refer this doc for more information on SLA https://azure.microsoft.com/en-us/pricing/details/api-management/

Regarding the alerts for this issue, you can implement something similar to this : https://learn.microsoft.com/en-us/azure/monitoring-and-diagnostics/monitoring-activity-log-alerts-on-service-notifications . With this implementation you should get a notification if the service is down.

However, you can prevent downtime by implementing a High Availability APIM deployed in multiple regions which would help you avoid any downtime: https://learn.microsoft.com/en-us/azure/api-management/api-management-howto-deploy-multi-region.

Please 'Accept as answer' and ‘Upvote’ if it helped so that it can help others in the community looking for help on similar topics.

Ankit Raj 1 Reputation point

2022-09-29T09:08:19.277+00:00

Hi @JananiRamesh-MSFT , Is there any way through which we can have two api management in developer mode and if one is down then traffic is redirected to another.
Our Current Infra is as follows:

I have a api management instance in developer tier in internal vnet mode.

We have application gateway in front of our api management to serve the traffic.

Our requirement:

Create an api management in developer mode in different region/country with same configuration as the previous one

Application gateway should be able to redirect traffic to both api management. Even if one is down in one region due to maintenance other should be able to serve the traffic

Thanks,
Ankit

JananiRamesh-MSFT 21,861 Reputation points

2022-10-06T13:27:36.177+00:00

Hi @Ankit Raj Thanks for getting back, Apologies for the delay in response. To answer your question, you can be able to create a APIM instance in two different regions with same configuration, but it is not possible to redirect traffic to another region if APIM in one region goes down as Application gateway is a regional load balancer, which means you cannot use it to redirect traffic to 2 different regions.

let me know incase of further queries, I would be happy to assist you.

Please 'Accept as answer' and ‘Upvote’ if it helped so that it can help others in the community looking for help on similar topics.
Sign in to comment